Studies on the connections between ... - Helda -

Studies on the connections between ... - Helda -


N:o 144 (2013)





Division of Atmospheric Sciences

Department of Physics

Faculty of Science

University of Helsinki

Helsinki, Finland

Academic dissertation

To be presented, with the permission of the Faculty of Science

of the University of Helsinki, for public criticism in auditorium E204,

Gustaf Hällströmin katu 2, on November 29 th , 2013, at 12 o’clock noon.

Helsinki 2013

Author’s address: Sanna-Liisa Sihto-Nissilä

Department of Physics

P.O. Box 64

FI-00014 University of Helsinki



Professor, Ph.D. Markku Kulmala

Division of Atmospheric Sciences

Department of Physics

University of Helsinki

Ph.D., docent Michael Boy

Division of Atmospheric Sciences

Department of Physics

University of Helsinki

Professor, Dr. Tech. Kari Lehtinen

Department of Physics, University of Eastern Finland, and

Finnish Meteorological Institute, Kuopio Unit

Kuopio, Finland


Ph.D., docent Harri Kokkola

Finnish Meteorological Institute, Kuopio Unit

Kuopio, Finland

Professor, Ph.D. Aijun Ding

School of Atmospheric Sciences

Nanjing University, China


Professor, Ph.D. Erik Swietlicki

Division of Nuclear Physics

Department of Physics

Lund University, Sweden

ISBN 978-952-5822-78-6 (printed version)

ISSN 0784-3496

Helsinki 2013

Unigrafia Oy

ISBN 978-952-5822-79-3 (electronic version)

Helsinki 2013

Helsingin yliopiston verkkojulkaisut


This thesis was made during 2004–2013 at the Division of Atmospheric Sciences of

University of Helsinki. I express my gratitude to Prof. Markku Kulmala, head of the

division, for the opportunity to work with such an interesting subject, atmospheric

science, in a motivated group. I thank the head of the Department of Physics, Prof.

Juhani Keinonen, for providing the working facilities needed for the research.

While located in the Department of Physics, the research at the Division of Atmospheric

Sciences is a fascinating multidisciplinary combination of physics, chemistry,

meteorology and even biology. I am grateful to have been able to work in this interesting

field, which in some aspects still resembles how natural sciences once started:

let’s go out to nature and see how it works! In our division people have mainly gone to

the forest in Hyytiälä and measured how new particles are formed there. During these

years, I have learned many interesting things — and many of them outside the basic

physics — for example which gases trees are emitting when they are stressed and how

water is transported in a tree.

I thank all my coauthors for good research co-operation. Especially I want to thank

Ilona Riipinen for excellent co-operation in preparing paper II, Hannele Korhonen for

giving her UHMA-code to my use, Henri Vuollekoski and Johannes Leppä for doing

modelling work together, and Joonas Vanhanen and Jyri Mikkilä for co-operation in

my last paper about CCN. This thesis combines modelling and analysis of field data. I

personally did not participate the measurements, and therefore I want to acknowledge

all researchers responsible for measurements as well as the technicians working at the

field stations for providing the data used in this thesis.

I thank my supervisors Kari Lehtinen and Michael Boy for guidance during the

whole path of PhD studies. Veli-Matti Kerminen deserves thanks for commenting

the manuscripts and for supervision when I was finalizing this thesis. I thank my reviewers

for the constructive comments to improve the introduction part as well as PhD

Theo Kurtén for proofreading the thesis. Maj and Tor Nessling Foundation (project no

2008310) and Academy of Finland (Center of Excellence Program) are acknowledged

for financial support.

I thank all the staff at the Division of Atmospheric Sciences for creating a nice working

environment andofferinggoodcompany. Beingnostalgic, Iwant tomentionthe”gang”

who started the doctoral studies at about the same time and/or with whom I shared

a room: Ilona Riipinen, Anne Hirsikko, Lauri Laakso, Tareq Hussein, Theo Kurtén,

Antti Lauri, Henri Vuollekoski, Tuomo Nieminen, Martta Toivola (Salonen), Johanna

Lauros and Eija Asmi (Vartiainen). Thank you for company in research, studies,

Hyytiälä courses and during conference trips, and for many many discussions both

about scientific and non-scientific topics.

This period of PhD studies coincided with a difficult period in my personal life, and

that certainly has affected my work too and made the period of PhD studies longer

thanexpected. Ithankmy supervisors andcoauthorsforpatience andhumane attitude

— we all are simply humans and face problems from time to time.

I express my gratitude for collaboration and friendship with PhD Amar Hamed, a

colleague from Kuopio, who passed away just two months ago. Thank you Amar for

these years.

Last I want to thank my family and relatives for support, especially during the difficult

times. Encouragements personally and via internet were important for finalizing

the thesis this spring/summer. I also have to give my acknowledments to music and

handcrafts, those are often needed to keep things in balance! I am most grateful to my

husband Jaani for support, patience, belief in my abilities even on bad days, and for

all kinds of practical help — this, I guess, may be called love.

In Kumpula, Helsinki, October 2013

Sanna-Liisa Sihto-Nissilä

ong>Studiesong> on the connections between atmospheric sulphuric acid, new particle

formation and cloud condensation nuclei

Sanna-Liisa Katariina Sihto-Nissilä

University of Helsinki, 2013


Atmosphericaerosol particles (small nmto µmsizedparticles floatingin air) areanimportant

part of the atmosphere and the climate system. Aerosols directly scatter sunlight and influence

cloud formation, thereby causing a net cooling effect on the climate which counteracts

global warming caused by greenhouse gases. Aerosols, particularly those from anthropogenic

pollution, also deteriorate human health.

Aerosols originate either from direct particle emissions or are formed in the atmosphere

from gas-phase vapours through nucleation. Aerosols have both anthropogenic and natural

sources. In the atmosphere, particle formation occurs frequently in continental areas all

around the globe, and it is an important source of aerosol particles and cloud condensation

nuclei. Sulphuric acid is one of the main compounds in atmospheric particle formation, and

it participates both in nucleation and particle growth.

This thesis studied the process of atmospheric particle formation and specifically its connection

to gaseous sulphuric acid, based on analysis of field measurement data and modelling.

New particle formation rates were observed to correlate with sulphuric acid concentration to

the power between 1–2. This correlation was notably different than expected on the basis of

nucleation theories. Based ontheobserved linear andsquaredcorrelation, newsemi-empirical

parameterisations for atmospheric nucleation rate were proposed: the activation and kinetic

nucleation mechanism. Empirical nucleation coefficients were determined from atmosperic

field data measured at two field stations in Hyytiälä, Finland, and Heidelberg, Germany.

The correlation of new particle formation with sulphuric acid and the factors affecting the

correlation were further investigated by performing simulations with an aerosol dynamical


Atmospheric relative humidity was observed to correlate negatively with sulphuric acid concentration

and particle formation rate. It was proposed that cloudiness at high relative humidities

could decrease the amount of UV-radiation reaching the ground, thereby decreasing

the formation of sulphuric acid through a photochemical reaction pathway.

Thisthesis also investigated the ability of aerosol particles in boreal forest to act as cloud condensation

nuclei (CCN). New particle formation events were observed to produce significant

amounts of potential CCN.

The results of this thesis provided new insights on atmospheric particle formation and its

connection to sulphuric acid. The developed nucleation rate parameterisations are useful for

modelling of aerosol formation in regional and global climate models. The CCN parameters

determined for boreal forest environment can be applied in climate modelling to predict

boreal forest aerosols’ effects on climate.

Keywords: atmospheric aerosol, particle formation, nucleation, sulphuric acid, cloud condensation



List of publications 7

1 Introduction 8

2 Atmospheric particle formation 13

2.1 Aerosol size distribution and its dynamics . . . . . . . . . . . . . . . . 15

2.1.1 Condensation and coagulation . . . . . . . . . . . . . . . . . . . 17

2.1.2 General dynamical equation . . . . . . . . . . . . . . . . . . . . 20

2.1.3 Condensation and Coagulation sinks . . . . . . . . . . . . . . . 21

2.2 Nucleation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

2.2.1 Basic concepts of nucleation theory . . . . . . . . . . . . . . . . 23

2.2.2 Atmospheric nucleation mechanisms . . . . . . . . . . . . . . . . 26

2.2.3 Laboratory measurements of atmospheric nucleation . . . . . . . 28

2.3 Formation and loss processes of sulphuric acid in the

atmosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

2.4 Activation of aerosol particles to cloud droplets . . . . . . . . . . . . . 34

3 Methods 37

3.1 Experimental data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

3.2 The calculation of particle formation rate . . . . . . . . . . . . . . . . . 41

3.2.1 Particle formation rate at 3 nm . . . . . . . . . . . . . . . . . . 42

3.2.2 Estimation of the nucleation rate from the apparent particle formation

rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

3.3 Evaluation of the calculation method of J 3 . . . . . . . . . . . . . . . . 47

3.4 University of Helsinki Multicomponent Aerosol model . . . . . . . . . . 50

4 Connection between sulphuric acid and new particle formation 53

4.1 General correlation of sulphuric acid and new particle formation in the

field data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

4.2 Activation and kinetic nucleation mechanisms . . . . . . . . . . . . . . 56

4.3 The effect of relative humidity on the nucleation rate . . . . . . . . . . 61

4.4 Modelling the connection between sulphuric acid and particle formation 65

5 CCN activity of boreal forest aerosols 69

5.1 Seasonal variation of CCN properties at the Hyytiälä

SMEAR II station . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

5.2 Effect of new particle formation on cloud condensation nuclei . . . . . . 72

6 Review of papers and the author’s contribution 75

7 Summary and conclusions 76

References 79

List of publications

This thesis consists of an introductory review, followed by six research articles. In the

introductory part, the papers are cited according to their roman numerals.

I Sihto, S.-L., Kulmala, M., Kerminen, V.-M., Dal Maso, M., Petäjä, T., Riipinen,

I., Korhonen, H., Arnold, F., Janson, R., Boy, M., Laaksonen, A. and Lehtinen,

K. E. J.: Atmospheric sulphuric acid and aerosol formation: implications

from atmospheric measurements for nucleation and early growth mechanisms,

Atmos. Chem. Phys., 6, 4079–4091, 2006.

II Riipinen, I., Sihto, S.-L., Kulmala, M., Arnold, F., Dal Maso, M., Birmili, W.,

Saarnio,K., Teinilä, K., Kerminen, V.-M., Laaksonen, A., Lehtinen, K.E.J.: Connections

between atmospheric sulphuric acid and new particle formation during

QUEST III–IV campaigns in Heidelberg and Hyytiälä, Atmos. Chem. Phys., 7,

1899–1914, 2007.

III Hamed, A., Korhonen, H., Sihto, S.-L., Joutsensaari, J., Järvinen, H., Petäjä,

T., Arnold, F., Nieminen, T., Kulmala, M., Smith, J.N., Lehtinen, K.E.J., and

Laaksonen, A.: The role of relative humidity in continental new particle formation,

J. Geophys. Res., 116, D03202, 2011.

IV Sihto, S.-L., Vuollekoski, H., Leppä, J., Riipinen, I., Kerminen, V.-M., Korhonen,

H., Lehtinen, K. E. J., Boy, M., and Kulmala, M.: Aerosol dynamics

simulations on the connection of sulphuric acid and new particle formation, Atmos.

Chem. Phys., 9, 2933–2947, 2009.

V Vuollekoski, H., Sihto, S.-L., Kerminen, V.-M., Kulmala, M., and Lehtinen, K.

E. J.: A numerical comparison of different methods for determining the particle

formation rate, Atmos. Chem. Phys., 12, 2289–2295, 2012.

VI Sihto, S.-L., Mikkilä, J., Vanhanen, J., Ehn, M., Liao, L., Lehtipalo, K., Aalto,

P.P., Duplissy, J., Petäjä, T., Kerminen, V.-M., Boy, M., and Kulmala, M.:

Seasonal variation of CCN concentrations and aerosol activation properties in

boreal forest, Atmos. Chem. Phys., 11, 13269–13285, 2011.

Papers I–II and IV-VI are reprinted under the Creative Commons Licence. Paper III

is reprinted with the permission of the journal.


1 Introduction

The air surrounding us is a mixture of gases (nitrogen, oxygen, argon, water vapour,

carbon dioxide etc.) and small particles. Air is an example of an aerosol: a mixture of

gas and small particles floating in it. The floating particles are called aerosol particles.

Due to their small size, in most cases aerosol particles are invisible to the human

eye. However, at sufficiently big particle sizes and number concentrations the particles

becomevisible, asseeninpollutionfromcarexhaustpipeorfromindustrysmokestacks,

as smog in polluted cities, dust storms in desert areas or haze in a moist forest. In

additiontotheatmosphere, aerosolsareencounteredinmanytechnicalapplications, for

example: in deodorant spray, paints and other liquids that can be dispersed smoothly

by first spraying them as aerosols to air or in inhalators used in drug delivery. This

thesis deals with aerosol particles in the atmosphere, i.e. atmospheric aerosols.

Atmospheric aerosols originate either from direct particle emissions (primary aerosols)

or are formed in the atmosphere from gases through nucleation (secondary aerosols).

The direct particle emissions include soot and other particle emissions from wood and

fossil fuel combustion, road dust particles, sea salt aerosols suspended to air from

whitecaps on the sea surface, pollen emitted from flowering plants and trees, bacteria

and viruses floating in air, and dust particles removed from Earth’s surface by blowing

wind (e.g. Reid et al., 2005; Engelstaedter et al., 2006; Viana et al., 2008; Hultin et al.,

2011; Wang et al., 2013). These processes generate aerosol particles at all particle

sizes from ∼10 nm upto several tens of µm. In atmospheric particle formation, new

particles areproduced when vapour molecules collide together andformastable cluster

of 1–2 nm size (Kulmala, 2003; Kulmala et al., 2007, 2013). This process is called

nucleation. In the atmosphere, sulphuric acid is considered as the most important

nucleating vapour.

After being emitted to or formed in the atmosphere, aerosol particles are subject to

various physical and chemical processes. Particles grow when different vapours condense

onto the particle surfaces. Due to their random Brownian motion in air, aerosol

particles collide with each other and stick or merge together in a process called coagulation.

Chemical reactions on the particle surface or inside the liquid phase change

the chemical composition of the particles. Finally, the particles are removed from the

atmosphere when they deposit on any available surfaces: on the ground, ocean and

lake surfaces, tree leaves, walls, windows, etc. In addition to dry deposition, particles

are washed out from atmosphere by falling rain droplets or snow flakes.

Atmospheric aerosols cover a wide range of particle sizes from molecular clusters close

to 1 nm to big dust or sea-salt particles of several µm in diameter. The upper size limit

is set by gravity, which makes large particles fall down rapidly. The aerosol number

concentration varies from a few 100 cm −3 in very clean Arctic areas (Koponen et al.,

2003) to several 100 000 cm −3 in polluted megacities (Mönkkönen et al., 2005; Wu

et al., 2008). The large range of variation is a challenge for instrumentation, as the

same instrument has to be capable of measuring three to four orders of magnitude

particle size range and even five orders of magnitude concentration range.


The lifetime of an aerosol particle in the lower atmosphere is on the order of one day

to one week (e.g. Williams et al., 2002). Despite their short lifetime, there is a persistent

aerosol population in the atmosphere, maintained by the continuous emission and

removal of aerosol particles into and out of the atmosphere. Due to different physical

processes — nucleation, particle growth by condensation, collisions between particles,

removal by gravitational settling or diffusion to surfaces — atmospheric aerosol has a

characteristic particle size distribution. Aerosol particles tend to center around some

particle sizes, thus forming different modes. Typically atmospheric aerosol has 3–4

modes: a nucleation mode around particle sizes ∼10–30 nm resulting from new particle

formation, an Aitken mode around 30–100 nm (named after a pioneer aerosol

scientist who first observed it), an accumulation mode around 0.1–1 µm, and a coarse

mode of large particles around 1–10 µm. The accumulation mode results from the

fact that for this size range the removal processes are slowest, making particles of this

size accumulate in the atmosphere (Laakso et al., 2003; Mammarella et al., 2011). In

addition, there is a persistent cluster mode at sizes 1–2.5 nm (Kulmala et al., 2007).

Due to their short lifetime, aerosol particles travel in the troposphere (lower part of the

atmosphere) typically some hundreds of kilometers, at maximum some thousands of

kilometers. Thus, in constrast to long-lived greenhouse gases, tropospheric aerosols are

not uniformly distributed around the globe, but are a local or regional phenomenon.

However, in the stratosphere aerosol particles (e.g. those emitted from volcanic eruptions)

survive much longer, having effects on weather and climate on longer time scales

(Deshler, 2008).

Locally, aerosols affect air quality and deteriorate visibility (Chang et al., 2009a; Wang

et al., 2010). Elevated particle mass (PM) concentrations have been observed to correlate

with increased mortality in big cities (Dockery and Pope, 1994; Brunekreef and

Holgate, 2002; Peters and Pope, 2002). Based on these studies, regulations for safe

particle mass concentrations in size ranges < 10 µm (PM 10 ) and < 2.5 µm (PM 2.5 )

have been set up. Recently it has appeared more and more evidence, that especially

fine particles (< 2.5 µm) have negative health effects (Pope et al., 2002). While the

large (> 2.5 µm) particles deposit already in the nose and upper respiratory track, fine

particles can be transported far in our respiratory system and some of the ultrafine

particles (< 0.1 µm) may even enter the blood circulation system. It is expected, that

in the future there will be regulations not only for the particle mass but also for the

particle number concentration.

Despite their local nature, aerosols have regional and global effects on the weather and

climate (Mitchell et al., 1995; Haywood and Boucher, 2000; Lohmann and Feichter,

2005; Bennartz et al., 2011). Aerosol particles directly scatter and absorb solar and

infrared radiation, thereby affecting the amount of radiation reaching or escaping from

the Earth’s surface (the direct effect of aerosols) (Yu et al., 2006). Indirectly, aerosols

affect cloud formation by acting as cloud condensation nuclei (CCN) (Twomey, 1991;

Penner et al., 2004). Aerosols are crucial for cloud formation, because in the Earth’s

atmosphere the water vapour does not form droplets alone. Cloud droplets are formed

when water vapour nucleates (condenses) heterogenously on the surface of an aerosol

particle — thus every cloud droplet has an aerosol particle inside it. The amount of


aerosol particles available for cloud droplet formation affects the cloud cover as well as

the cloud properties: if there are more aerosols, there will be more but smaller cloud

droplets, as the same amount of water is divided to a larger number of aerosol particles.

Smaller cloud droplets reflect sunlight more efficiently i.e. make whiter clouds (first

indirect effect). Also, smaller cloud droplets are less eager to fall down as rain, and the

lifetime of the cloud will increase (second indirect effect).

From a climate point-of-view, the quantity of interest is the aerosols’ effect on the

radiative balance of the Earth (see Fig. 1). It is estimated that both the direct and

indirect effect have a total negative effect on the radiative balance (negative radiative

forcing): the more aerosols, the more sunlight will be reflected from aerosol particles

and clouds back to space. The absorption of sunlight and infrared radiation by aerosol

particles (especially black carbon aerosols) and cloud droplets (especially in high-level

clouds) causes a small warming effect, but in total the aerosol effect is estimated to be

cooling. TheIntergovernmental PanelonClimateChangereportgivesanestimate−0.5

W/m 2 for the direct and −0.7 W/m 2 for the indirect effect (IPCC, 2007). Overall, the

estimate for the total aerosol cooling effect (−1.2 W/m 2 ) is comparable to the warming

effect of CO 2 (1.66 W/m 2 ). However, the aerosol effect is associated with the largest

error bars (see Fig. 1), and the level of scientific understanding is stated to be low, in

contrast to greenhouse gases, whose effect is rather well understood. The research in

this thesis is one small contribution to quantify the aerosol effect more precisely, and

to reduce the uncertainty related to it.

In recent 20–30 years, considerable decreases in anthropogenic aerosol emissions have

happened in the Western countries, due to cleaning and filtering of the gas exhausts of

industry and transport systems (Hamed et al., 2010; Asmi et al., 2013). In the future,

this trend will continue, thus decreasing the man-influenced aerosol cooling effect. In

view of climate change, the major environmental problem of our time (Archer and

Pierrehumbert, 2011), this raises a concerning question: If the aerosol cooling effect

has been large, how much will the warming be accelerated, when the aerosol cooling

effect continues to decrease (Andreae et al., 2005; Arneth et al., 2009)? It is clear that

aerosols need to be taken into account in climate models when making predictions on

future climate change.

Measurements all around the world have shown that new particle formation seems

to occur almost everywhere on the Earth’s land-surface where aerosol measurement

instruments have been carried to (Kulmala et al., 2004d; Kulmala and Kerminen,

2008; Vakkari et al., 2011; Kyrö et al., 2013). Sulphuric acid has been identified as a

key compound in atmospheric particle formation and it participates both in nucleation

and particle growth. However, equally important for natural particle formation are

organic vapours, which contribute to particle condensational growth and account for

the major part of atmospheric aerosol mass (Jimenez et al., 2009; Riipinen et al.,

2011). The atmospheric aerosol is to some extent a self-regulating system: if there are

many particles emitted from other sources, such as pollution, there are less particles

produced by new particle formation; on the other hand, if the atmosphere is very clean

of particles, new particle formation occurs more frequently. Thus, atmospheric particle

formation maintains the aerosol particle population in the atmosphere.


Figure 1: Estimates for radiative forcing (RF) components for year 2005 (global averages).

Positive radiative forcing means a warming effect on climate and negative

radiative forcing a cooling effect. From IPCC (Intergovernmental Panel on Climate

Change) Fourth Assessment Report (IPCC, 2007).

Inoceanic areas andinrain forests inAmazonian, new particleformationin theboundary

layer occurs very rarely. There are still plenty of aerosols from other sources in

these environments, such as: sea salt particles and particles transported from the upper

atmosphere to the boundary layer on the oceans, biomass burning aerosols and

secondary organic aerosols in the Amazonian region, as well as particles from anthropogenic


One way to characterize aerosols is whether they are of natural or anthropogenic origin.

For primary aerosols, the origin is easily identified: sea salt, sand, pollen, bacteria,

viruses and other micro-organisms are examples of natural aerosols while all kinds of

pollution from combustion processes and biomass burning are examples of anthropogenic

aerosols. For secondary aerosols, the situation is not that easy. At first sight,

new particle formation could be called as natural, as it is happening ”naturally” in

the air — but new particle formation is to a large extent controlled by sulphuric acid,

which has both natural and anthropogenic sources. Sulphuric acid is formed in the atmosphere

mainly by oxidation of SO 2 , for which a major source is the burning of fossil

fuels. Volatile organic compounds, which condense on aerosol particles and constitute

most of the secondary aerosol mass, are mainly emitted by vegetation, but have also

anthropogenic sources (combustion of fossil fuels and other organicmatter). Thus, natural

and anthropogenic sources of particles and condensing vapours are mixed together


so that it is not fully possible to distinguish them. Also some natural primary emissions

are influenced by human activity: for example desertification increases emission

of mineral dust into air.

This thesis investigates atmospheric particleformationintheborealforest environment

by the methods of field data analysis and aerosol dynamical simulations. The main

emphasis is on studying the correlation of new particle formation with sulphuric acid

concentration. More specifically, this thesis aims to answer to the following research

questions or objectives:

• How arenew particle formationandgas phase sulphuric acid connected witheach


• To develop empirical parameterisations for the nucleation rate. What are the

values of the empirical nucleation coefficients indifferent environments and which

parameters do they depend on?

• Which factors, besides the nucleation mechanism, affect the correlation of the

particle formation rate with sulphuric acid?

• How accurate is the method used to calculate particle formation rate from the

size distribution data?

• What is the ability of aerosol particles in the boreal forest environment to act

as cloud condensation nuclei (CCN), and is new particle formation affecting the

CCN concentrations?

These research questions are discussed in this introductory part and in the six articles

included in this thesis.


2 Atmospheric particle formation

Inatmosphericparticleformation, newnm-sizedparticlesareformedintheatmosphere

from precursor vapours. The process is initiated by nucleation, in which small stable

clusters, of size 1–1.5 nm in ”diameter”, are formed. The precursor vapours must

be ”condensable”, i.e. they must have low saturation vapour pressures, so that the

vapours are eager to nucleate and stay in the condensed phase. The surrounding inert

gas is air, and air molecules do not participate in nucleation. The formed clusters grow

further by condensation of different vapours, eventually reaching sizes of 100–500 nm,

unless scavenged by the various aerosol removal processes. The upper size limit of

aerosol particles in atmosphere is set by gravitation: particles bigger than a few tens

of µm fall down rapidly due to gravitation.

Atmosphericparticleformationtypicallyoccursaroundmidday, onsunnydayswithlow

background aerosol concentration (W. Birmili and A. Wiedensohler, 2000; Boy et al.,

2003; Birmili et al., 2003; Stanier et al., 2004; Lyubovtseva et al., 2005; McMurry et al.,

2005). Whileparticleformationfromgasphasecompounds (gas-to-particleconversion)

has been known to take place, and to be a source of new particles in the atmosphere,

already for a long time, it was only in late 1990’s that the whole process of atmospheric

particle formation was recorded for the first time (Weber et al., 1996; Mäkelä et al.,

1997). Since 1997, the continuous measurements of particle size distributions at the

Hyytiälä forestry field station (SMEAR II), from 3 nm to 500 nm and with 10 min.

time resolution, have shown that in boreal forest conditions new particle formation

occurs frequently all around the year, on 50–120 days per year (Aalto et al., 2001; Dal

Maso et al., 2005).

Figure 2 shows a typical example of a new particle formation event, measured by a

Differential Mobility Particle Sizer (DMPS) at the SMEAR II station in Hyytiälä,

Finland. Nucleation produces new 1–2 nm sized particles from gas-phase precursor

vapours, below the detection limit (diameter 3 nm) of the DMPS. After nucleation,

the clusters grow in size by condensation of vapours; these vapours can be the same or

different than the nucleating vapours. Due to their random Brownian motion in air,

the particles collide with each other and stick together, in a process called coagulation.

Due to condensation and coagulation, the new nucleation mode shifts to larger particle

sizes. When particles reach the size range > 50 nm (in diameter), they start to have

effects on climate: the particles can act as cloud condensation nuclei (CCN) i.e. be

seed particles for cloud droplets (Kerminen et al., 2012). The aerosol particles are

subject to these dynamical processes, until they are removed from the atmosphere by

gravitational settling and diffusion onto surfaces (dry deposition), or are scavenged by

falling rain droplets (wet deposition).

In the atmosphere, there is always a background aerosol distribution present. Often

simultaneously to or just before the start of nucleation, this background particle concentration

is decreasing due to the turbulent mixing initiated by sunlight warming the

ground in the morning. The onset of turbulent mixing increases the boundary layer

height (boundary layer = the lowest, turbulently mixed layer of the atmosphere), thus


Hyytiälä March 25 th 2003



Diameter (m)

10 −7


10 −8




0:00 06:00 12:00 18:00 24:00

Time (hours)

10 100 1000 10000 100000

Concentration dN/dlog(d p

) (cm −3 )

Figure 2: An example of a new particle formation event measured by DMPS (Differential

Mobility Particle Sizer) at the SMEAR II station in Hyytiälä, Finland. The

surface plot presents the evolution of the particle size distribution with time on the

x-axis, particle diameter on the y-axis (log scale), and the normalized number concentration

indicated by the colour code. The aerosol dynamical processes modifying the

size distribution (at an exemplary location on diameter and time axes) are indicated.

mixing the particle-rich airmass of the boundary layer, left from the previous day, with

an upper airmass having a much lower particle concentration. This results in a remarkabledilution

of theparticle concentrations inthe morning hours (see Fig. 2). There are

observations that new particle formation is enhanced in conditions of strong turbulent

mixing (Nilsson et al., 2001; Lauros et al., 2007). However, it is not clear whether

the enhancement is due to turbulence itself or due to the decrease of the background

aerosol concentration.

In continental areas, large amount of evidence shows that sulphuric acid (H 2 SO 4 ) participates

in both formation and growth of new particles. Sulphuric acid is thought

to be main nucleating compound, together with water and ammonia (see Sect. 2.2),

although some organic molecules may also participate in nucleation. Of the particle

growth(bycondensationofvapours)sulphuricacidcanexplainonlyaminorpart, while

the main part is accounted for by various organic vapours present in the atmosphere.

For example, in boreal forest conditions in Hyytiälä, the sulphuric acid contribution to

the observed particle growth rate (determined as GR = dd p /dt, where d p is the particle

diameter) is estimated to be 8–30 % (Boy et al., 2005). In polluted cities, where SO 2

and H 2 SO 4 concentrations are high, the sulphuric acid contribution is higher (34–65 %),

but even in those conditions it does not alone explain the particle growth (Yue et al.,

2010; Gao et al., 2011).


Theparticlegrowthrates, determinedfromtheparticlesizedistributionmeasurements,

show a seasonal variation according to the growth season of the vegetation, implying

that organic compounds emitted by vegetation are important for particle growth (Hirsikko

et al., 2005; Yli-Juuti et al., 2011). A major source for condensable organic

vapours in the atmosphere are volatile organic compounds (VOCs) emitted by vegetation,

although anthropogenic VOC emissions also exist. The main biogenic VOC

groups are isoprene, monoterpenes and sesquiterpenes. In boreal forest areas, monoterpenes

(chemical formula C 5 H 10 , having several different molecular structures) emitted

from coniferous trees constitute the main part of VOC emissions (Haapanala et al.,

2007), while in deciduous forests isoprene is dominating (Carlton et al., 2009). The

VOCs do not condense directly on particles, since they are volatile i.e. have a high

saturation vapour pressure, but when VOCs are oxidised in the atmosphere by OH,

O 3 and NO 3 , semi-volatile reaction products having a lower saturation vapour pressure

are formed. The VOC oxidation chemistry is complex and the exact identities of

the condensable vapours are still unknown. However, these organic compounds have

recently been observed directly by in situ measurements of aerosol particle chemical

composition, showing that a major part of the aerosol mass is organic (Jimenez et al.,

2009; Laitinen et al., 2011; Zhang et al., 2011).

As a simple approximation, in aerosol science the formed particles are typically considered

to be liquid spheres, and mostly the modelling of aerosol formation is based on

this assumption. However, for very small clusters close to 1 nm one cannot accurately

determine the phase and diameter, and it would rather be more correct to speak about

a molecular cluster than a particle. From 3 nm upwards the number of molecules in

the particle is high enough, that we can speak of a ”macroscopic” particle, with a well

defined diameter and phase. For particles larger than 3 nm the assumption of spherical

liquid particles is usually good (excluding agglomerates of e.g. soot particles), based

on the typical temperatures in atmosphere and the fact that particles are growing

mainly due to condensation, thus making a liquid phase particle. However, there are

new observations which indicate that the phase of the formed particles is amorphous

(Virtanen et al., 2010), and it might be that the phase of aerosol particles needs to be

reconsidered in more detail in future studies.

In recent years, great advances in aerosol measurement technology have been achieved,

making it possible to detect the process of atmospheric nucleation directly from the

nucleation size at 1–2 nm (Zhao et al., 2010; Kulmala et al., 2012, 2013). However,

the details of atmospheric particle formation and particle growth are still not fully


2.1 Aerosol size distribution and its dynamics

The atmospheric aerosol particle population is described by the particle concentration,

size (diameter) and chemical composition. Typically, in physical studies of aerosols,

the chemical composition is neglected, and the aerosol is characterized by an aerosol

size distribution function, expressed either for the number concentration as the particle


number size distribution (mathematically denoted as dN(d p )/dd p or dN(d p )/dlog(d p ))

or for the mass concentration as the particle mass size distribution (dm(d p )/dd p ). The

atmospheric particle number size distribution peaks at small particle sizes, while the

mass distribution peaks at larger particle sizes as mass is proportional to the particle

volume (∼ d 3 p ) (see Fig. 3). The actual number or mass concentration (for a certain

size range and having units 1/cm 3 or µg/cm 3 ), is obtained by integration over the

desired size range.

In aerosol dynamical studies, most often we consider the number size distribution.

However, in air quality studies and the regulation standards for particulate pollution,

the mass concentration is used. In the following, the shortened term particle size

distribution means the particle number size distribution and the number concentration

has units 1/cm 3 . Due to the log-normal distribution with respect to particle size,

number concentration is most oftenexpressed inthe normalized formdN(d p )/dlog(d p ).


x 10 −17

dN/dlog(d p

) (1/cm 3 )






10 −9 10 −8 10 −7 10 −6 10 −5

Particle diameter (m)




dm/dlog(d p

) (µg/cm 3 )

Figure 3: An average particle number size distribution (red) and the corresponding

mass size distribution (black) in Kumpula, Helsinki, in springtime (Hussein et al.,

2004). The number size distribution consists of three modes, which are indicated by

dashed lines; for the mass distribution (assuming particle density of 1 g/cm 3 ) only the

total distribution is plotted.

The atmospheric aerosol size distribution is typically composed of 2–4 log-normally

distributed modes: a nucleation mode at < 30 nm, an Aitken mode at 30–100 nm,

an accumulation mode at 100 nm–1 µm and a coarse mode at 1–10 µm. Aitken and

accumulationmodesarealwayspresent intheatmosphere, withtheaccumulationmode

being the most persistent, because for that size range the removal mechanisms (dry

and wet deposition) are the slowest. The nucleation mode emerges on new particle

formation days and the strength of the coarse mode is dependent on primary emissions

of big particles, such as dust emissions. The atmospheric aerosol distributions differ in

different environments: in continental areas the distribution has 3–4 modes, whereas in


marine areas, where nucleation is rare, thedistribution isbi- or trimodal (e.g. Koponen

et al., 2002).

The aerosol size distribution is changing all the time due to various aerosol dynamical

processes, which modify particle concentrations, size and chemical composition. The

particles collide with each other and stick together, forming bigger particles or agglomerates,

in the process called coagulation. Particles grow by condensation of vapours

onto particle surfaces or shrink by evaporation of molecules from particle surfaces to

the gas phase. Nucleation inserts new, nm-sized particles into atmosphere, when condensable

vapours form new stable particles. Particles are removed from atmosphere

by dry and wet deposition: in dry deposition particles diffuse and stick to macroscopic

surfaces, e.g. tree leaves, or settle down to the Earth surface by gravitation; in wet

deposition particles are washed away from atmosphere as they collide with falling rain

droplets or with cloud droplets.

The processes mentioned above are the main aerosol dynamical processes acting in the

boundary layer. In clouds, aerosols may also undergo changes in size and chemical

composition. Every cloud droplet has an aerosol particle inside it (e.g. McFiggans

et al., 2006). Vapours and gases present in the cloud (such as nitric acid, sulphur

dioxide, organic compounds and water) condense or dissolve to particles, and after that

the compounds undergo chemical reactions in the aqueous phase. When the prevailing

ambient conditions change, some amount of condensed material may evaporate from

the particle, leaving the seed particle composition and size changed in comparison to

the original state (e.g. Romakkaniemi et al., 2006). This is called cloud processing

of aerosol particles. A special bimodal structure of aerosol size distributions, often

encountered in marine boundary layer, is related to several cloud processing cycles of

aerosol particles (Hoppel et al., 1996).

In addition to these physical processes, there may be surface reactions or heterogenous

nucleation (nucleation on top of an existing particle) that change the particle size and

composition. Chemical reactions and polymerisation processes inside the particle can

modify the chemical composition and stability of the aerosol particle (aging of aerosol).

2.1.1 Condensation and coagulation

Because the size distribution of aerosol particles spans from the molecular scale (diameter

of few nm) to microscopic scale (diameter of few µm), different theoretical frameworks

are needed: small nm-sized particles are in the free-molecular regime, where

particles behave similarly to gas molecules, whereas large particles are in the continuum

regime, where particle experiences the surrounding gas as a continuous fluid. In

the continuum regime, both condensation and coagulation in aerosol systems are dealt

with the aid of diffusion theory. With condensation the use of diffusion theory is selfevident,

but also coagulation in the continuum regime can be considered as a diffusion

process: small particles in a fluid are diffusing towards a big (stationary) particle. In

the free molecular regime, the kinetic gas theory applies both for condensation and


coagulation. In between these, there is a transition regime, in which both diffusive and

free-molecular effects are important.

The condensation or evaporation flux is governed by the equilibrium vapour pressure

(or concentration) prevailing at the particle surface. On the curved particle surface,

the equilibrium pressure is always higher than on a planar surface of the same composition,

because of weaker atomic bonding between surface molecules/atoms on a curved

surface. This curvature effect on the equilibrium (saturation) vapour pressure is called

the Kelvin effect and it is determined by the Kelvin equation:

( 4Mv σ


p eq = p sat exp , (1)

RTρd p

where p sat is the vapour saturation pressure for a planar surface, M v is the molar mass

of the vapour, σ is the surface tension of the liquid, R is the universal gas constant,

T is temperature, ρ is the density of the liquid and d p is the diameter of the liquid

particle (Seinfeld and Pandis, 2006; Vehkamäki, 2006).

The difference between the vapour concentration far from the particle and the equilibrium

concentration at the particle surface determines the direction and magnitude of

the mass flux towards/from the particle (I m ∝ (c vapour −c eq )). The Kelvin effect limits

the condensation of vapours on a curved surface: the smaller the particle diameter, the

more eagerly the molecules evaporate from the particle surface.

Due to the Kelvin effect, the condensation onto small, few-nm sized particles, would

be extremely difficult for most of the condensable vapours (e.g. oxidation products of

VOCs) present in the atmosphere. Sulphuric acid is the exeption: it has a very small

saturation vapour pressure (< 10 −4 Pa, Ayers et al., 1980; Marti et al., 1997), so that

even with the Kelvin effect the saturation vapour pressure remains negligible, and it

starts to condense on particles directly after nucleation. However, also for the smallest

particles, the particle growth rate is explained only partly by sulphuric acid, and the

rest of the growth is attributed to various organic compounds (Riipinen et al., 2012).

To theoretically explain the organic vapour contribution on the particle growth rates,

Kulmalaetal.(2004b)presentedthenano-Köhlertheory, whichwouldfacilitateorganic

vapourcondensationonsmall(d p

where p sat,org is the organic vapour saturation pressure, γ org is the activity coefficient

of the organic substance, x org its molar fraction, M org the molecular mass and ρ org

the density. This mechanism for organic vapour condensation was applied in aerosol

dynamical simulations of paper III.

In the atmosphere, coagulation is caused by random Brownian motion of aerosol particles.

Due to velocity differences, both in direction and magnitude, particles collide

with each other. For aerosols, the collisions are assumed to be non-elastic, i.e. particles

alwayssticktogether. Often, thecollisionisassumedtobebetween liquidspheres, making

the particles merge together, thus forming a new, slightly larger spherical particle

(agglomeration is not taken into account). The collision frequency function (= coagulation

coefficient) can be derived from diffusion theory for the continuum regime and

from kinetic gas theory for the free molecular regime particles, and it depends strongly

on particle sizes. Coagulation is most efficient between particles of big size difference:

the large particle surface collects the small, rapidly moving particles. Therefore, coagulational

scavenging to backgound aerosol (at around 100–500 nm) is the main loss

mechanism for small, nucleation mode particles. The coagulation coefficient is smallest

and rather independent on particle size for the particles of equal size (Seinfeld and

Pandis, 2006).

In addition to Brownian motion, velocity differences resulting from gradients in air

flow, electrical or gravitational force field can cause collisions between particles. These

processes, however, are important only at high velocity or force gradients occuring e.g.

in flow tubes and are not significant coagulation processes in the atmosphere.

Thetheoryregime forcondensation andcoagulationisdefined bytheKnudsen number,

which describes the nature of the suspending fluid relative to the particle:

Kn = 2λ air

d p

, (3)

where λ air is the mean free path of air molecules (about 65 nm at T = 298 K and p

= 1 atm) and d p is the particle diameter. When Kn >> 1, aerosol particles experience

collisions similarly as molecules and we are in the kinetic regime. In this regime,

the kinetic gas theory applies. When Kn

With coagulation, thequantities to be compared arethe mean free path of thediffusing

aerosol particle (smaller particle, λ p ) and the diameter of the absorbing particle d p .

However, the mean free path of aerosol particles is only weakly dependent on particle

size, and the value is quite close to the air mean free path (λ p varies between 10–60 nm;

Seinfeld andPandis (2006, Table 8.5)). Because ofthat, thesamedefinition ofKnudsen

number (Eq. 3) can be used also in the case of coagulation to roughly characterize the

theory regime. In exact calculations, the different equations for transition regime and

free molecular coagulation coefficient use slightly different expressions for the Knudsen

number (Seinfeld and Pandis, 2006). The most widely used formula for the coagulation

coefficent (which is also used in this thesis) is the one by Fuchs (1964).

2.1.2 General dynamical equation


containing terms for all aerosol dynamical processes acting on the aerosol population.

The general dynamic equation (GDE), expressed in the volume space v = 1 6 πd3 p, reads

(Seinfeld and Pandis, 2006):



= 1 2

∫ v


K(v −q,q)n(v −q,t)n(q,t)dq−n(v,t)

∫ ∞



− ∂

∂v [I(v)n(v,t)]+J nucδ(v −v nuc )+S(v)−R(v), (4)

where n(v,t) = ∂N(v,t)/∂v is the particle size distribution function, t is time and

v is the particle volume, K(v,q) is the coagulation coefficient between particles of

volume v and q, I(v) is the total volume flux of vapour molecules onto the particle

due to condensation/evaporation, J nuc is the nucleation rate, v nuc is the volume of

the nucleated particle and δ(v − v nuc ) is the delta function, with a value of unity for

v = v nuc and otherwise zero. The first two terms on the right hand side represent

coagulation (production of v-sized particles in collisions of smaller particles and the

coagulationallossofv-sizedparticlesincollisionswithbackgroundparticles). Thethird

termrepresents condensational growthorshrinkageduetoevaporation, thefourthterm

represents nucleation, and S and R are possible additional source and removal terms.

In practice, the general dynamic equation is often used in a discrete form, as we deal

with measured discrete spectrums or utilize models with a sectional representation for

aerosol size distribution. In modal models, which express the aerosol size distribution

as a superposition of continuous log-normal modes, the GDE is used in the continuous

form of Eq. 4. In models, the GDE is integrated numerically to solve the particle size

distribution evolution. In aerosol studies, particles are often assumed to be spherical,

and the GDE can then be expressed more conveniently with the particle diameter.


2.1.3 Condensation and Coagulation sinks

Two useful quantities – the condensation sink and the coagulation sink – have been

introduced to characterize the aerosol size distribution in terms of condensation and

coagulation with one scalar quantity (Kulmala et al., 2001a).

The condensation sink (CS) describes the condensation rate of vapour onto the whole

particle size distribution:

vapour loss rate due to condensation = CS ×C v , (5)

where C v is the concentration of condensable vapour, i.e. the condensation sink is the

vapour condensation rate per one molecule and has units 1/s.

The equation for the condensation sink can be derived from condensation theory. It

depends on the background particle size distribution (particle surface area) and vapour

diffusion properties (Kulmala et al., 2001a):

∫ ∞

CS = 2πD v β m (d p )d p n(d p )dd p

∼ = 2πDv β m (d p,i )d p,i N i , (6)


where D v is the diffusion constant of the vapour, n(d p ) is the particle size distribution

function, d p is the particle diameter, and the integration is performed over the whole

particle size distribution. The latter form is for a discrete particle size distribution

N i (d p,i ). Because thediffusionconstant isvapour-specific, theCS hastobedetermined

for a specific vapour, typically for sulphuric acid (H 2 SO 4 ). The parameter β m (d p ) is the

Fuchs-Sutugin correction factor forthe transition regime mass flux (Fuchs andSutugin,



β m =


1+( 4

3α m

+0.377)Kn+ 4

3α m

Kn2, (7)

where Kn is the Knudsen number and α m is the mass accommodation coefficient.

With the Fuchs-Sutugin transitional regime correction factor, Equation 6 is valid for

allparticle sizes, fromkinetic to continuum regime; the semi-empirical correction factor

takes into account the changes in the condensation theory when particle size decreases.

The mass accommodation coefficient α m is also called ”a sticking coefficient” and it

describes the probability that a molecule, when hitting a particle surface, sticks onto

it. There has been quite much debate about the value of α m (Jefferson et al., 1997).

In most cases, it is assumed to be unity (α m = 1), meaning that when a molecule hits

a surface, it will be absorbed.

Analogously to the condensation sink, the coagulation sink is defined for a certain

particle size as the loss rate due to coagulation per one particle. For a discrete size

distribution this is expressed as:


Coagulation rate of particles of diameter d i = CoagS di ×N i ,

where CoagS di is the coagulation sink (unit 1/s) and N i is the number concentration

of particles in a size bin around particle diameter d i . The formula for coagulation sink

can bederived fromthe equation for Brownian coagulationrate, giving (Kulmala et al.,


CoagS(d p,i ,t) =

∫ ∞

d p,0

β(d p,i ,d ′ p)n(d ′ p,t)dd ′ p ∼ = ∑ j

β(d i ,d j )N j , (8)

where β(d p,i ,d ′ p) is the Brownian coagulation coefficient i.e. the collision frequency

function between particles of diameters d p,i and d ′ p (Seinfeld and Pandis, 2006). The

sum expression on the right is the formula for discrete size distribution.

Typically the coagulation sink is calculated for small nucleation mode particles, as


for nucleation mode particles. The coagulation rate and thus CoagS is largest for the

smallest particles, and decreases as the particle size increases (Dal Maso et al., 2002).

For small particles of d p = 1 nm the coagulationsink approaches the condensation sink,

because condensation can be viewed as collisions (= coagulation) of H 2 SO 4 molecules

(having ”diameter” below 1 nm) with background particles.

Some researchers have used the concept of Fuchs surface area instead of condensation


A Fuchs = 4π 3

∫ ∞


Knβ m (d p )d 2 pn(d p )dd p , (9)

the meaning is analogous to the condensation sink (McMurry and Friedlander, 1978;

McMurry et al., 2005). Connecting this with the equation for CS (Eq. 6) and applying

λ v = 3D v /¯c v for the mean free path in Kn (Seinfeld and Pandis, 2006), a relationship

between the condensation sink and the Fuchs surface area A Fuchs can be derived:

CS = 1 4 ¯c vA Fuchs , (10)

where ¯c v is the mean thermal velocity of the condensing vapour molecule.

The condensation and coagulation sinks are useful quantities to descibe the particle

size distribution with a one scalar quantity. Both are measures of total aerosol surface

area, the former in the view of condensation and the latter in view of coagulation. The

value of CS and CoagS is determined especially by the concentration of large particles,

which have a large surface area.


In new particle formation studies, CS is often used instead of CoagS, even if one

actually, inaconceptualsense, isreferringbothtothecondensationandthecoagulation

sink. Thephrase”newparticleformationeventsoccurpreferablyatlowCSconditions”,

actually means that both the CS and CoagS are low, having two effects favouring

nucleation: i) there will be more vapour available for nucleation and condensational


with large particles will be smaller (low CoagS).

2.2 Nucleation

2.2.1 Basic concepts of nucleation theory

By definition, nucleation is the first step of a phase transition process (Vehkamäki,

2006). All phase transitions start with nucleation. For example, freezing of water

typically starts with nucleation of small ice crystals around impurities present in water

(molecules, small particles). These crystals then grow in size and result in macroscopic

freezing of the water, if the temperature stays below zero. Similarly, boiling of water

begins with bubble formation (= nucleation of gas phase bubbles from liquid phase)

around impurities dissolved in water or present on the inner surface of the stewpot.

These are examples of heterogenous nucleation, in which nucleation happens on top of

a foreign surface, such as the surface of a small impurity particle. In distilled water,

which is free from impurities, nucleation (such as formation of ice crystals or bubbles)

happens homogenously without the aid of an exisiting surface. Homogenous nucleation

isalways energetically moredifficult thanheterogenous nucleation; therefore freezing of

distilled water requires lower temperatures and boiling happens at higher temperatures

than for normal water with impurities.

In atmospheric particle formation, we consider the transition from gas phase to liquid

or solid phase, i.e. the nucleation of liquid or solid phase clusters from gas-phase

precursors. Typically, for simplicity, the phase of the formed cluster or particle is

considered to be liquid.

As all physical processes in nature, nucleation is governed by energy. For nucleation

to happen, the energy state of a nucleated, liquid phase particle must be lower than

the initial state of vapour molecules. The first requirement for this is, that the vapour

is supersaturated. The saturation ratio S i (for vapour i) is defined as:

S i = p i,v

p i,sat

, (11)

where p i,v is the partial pressure of vapour i (in air) and p i,sat is its saturation vapour

pressure. The vapour is supersaturated when S i > 1, meaning that there is an excess

amount of molecules in the vapour phase and the liquid state would be energetically

more favourable.


Innucleation, in between the initial (vapour) andfinal (nucleated cluster) energy states

there is an energy barrier which needs to be crossed with the aid of thermal energy.

In classical nucleation theory, the probability of crossing the barrier is given by the

Boltzmann factor e −∆G/k BT (∆G is the height of the energy barrier, k B is Boltzmann’s

constant and T is temperature), and the nucleation rate is:

J nuc = K kin e −∆G∗ /k B T . (12)

In this equation ∆G ∗ is the Gibbs free energy of the formation of the critical cluster (=

height of the energy barrier) and K kin is a kinetic prefactor accounting for the collision

rate of vapour molecules with the cluster, which make the cluster grow. According

to classical nucleation theory the Gibbs free energy (for homogenous nucleation) is

(Seinfeld and Pandis, 2006; Vehkamäki, 2006):

∆G = −nk B T lnS +4πr 2 σ, (13)

where n is the number of molecules in the cluster, r is the cluster radius, S is the

saturation ratio of the vapour and σ is the surface tension. The first term represents

the gain in energy that is obtained in forming a liquid phase cluster, and the second

term is the energy needed to create a new surface (due to surface tension). The typical

form of the Gibbs free energy curve is shown in Figure 4, where the maximum of the

curve represents the critical point: the smallest stable cluster i.e. the critical cluster,

that does not tend to evaporate (radius r ∗ at saturation ratio S ∗ ). The clusters bigger

than r ∗ start to grow spontaneously by condensation, as the cluster moves downhill on

theGibbsfreeenergycurve(providedthatthesaturationratiostaysthesame). Typical

critical cluster sizes are ∼1–5 nm, number of molecules being from a few molecules to

∼100. In multicomponent nucleation, involving more than one compound, the Gibbs

free energy becomes a surface with i dimensions, i being the number of compounds.

Depending on the substances, multicomponent nucleation can be either easier or more

difficult than homogenous nucleation of the participating substances.

In deriving the equation for ∆G, several approximations were made. For example, the

cluster is assumed to be a liquid sphere, and it is assumed to have the properties (density,

surface tension) of bulk liquid. It is clear that these approximations are very rough

and do not hold very well for clusters of a couple of nanometers in diameter (Merikanto

et al., 2007). Despite its deficiences, classical nucleation theory is so far the best concise

theory for nucleation and it is useful in interpretation of experimental studies on

nucleation. In many cases the classical nucleation theory predicts the S-dependence

right but T-dependence wrong, suggesting that the theory predicts the size of critical

cluster correctly, but fails in describing the energy of the cluster (Vehkamäki, 2006).

For molecular clusters, ab initio quantum chemical calculations provide physically and

chemically more accurate description, and can give insights on the structure of nucleated

clusters (Kurtén et al., 2007; Torpo et al., 2007; Kurtén et al., 2008; Ortega et al.,



S = 0.9


S = 1.0

Gibbs free energy (J)





1 x 10−16 Radius (m)


S = 1.1

S = 1.15


0 0.5 1 1.5 2

x 10 −8

Figure 4: The Gibbs free energy curves for homogenous nucleation of water vapour at

four different saturation ratios. The Gibbs free energy has a maximum and nucleation

is possible when S > 1. The critical cluster size r ∗ and the corresponding energy of

the critical cluster (G ∗ ) at the maximum of the ∆G curve are indicated.

One of the most useful theoretical results for nucleation is the nucleation theorem. The

first nucleation theorem states, that the derivative of the logarithm of nucleation rate

with respect to the logatrithm of saturation ratio is related to the number of molecules

in the critical cluster (Vehkamäki, 2006):

∂lnJ nuc

∂lnS i

= n ∗ i +1 ≈ n∗ i , (14)

where J nuc is the nucleation rate, S i is the saturation ratio of the nucleating vapour,

and n ∗ i is the number of molecules of this substance in the critical cluster. In multicomponent

nucleation, the nucleation theorem applies separately for each nucleating


The first nucleation theorem is a general result and not restricted to any special nucleation

theory; thus it is valid more widely than the classical nucleation theory. In

principle, it applies for any nucleation process which has an energy barrier associated

with it. Therefore, it is very useful in interpretation of nucleation experiments. By

measuring the nucleation rate (J nuc ) as a function of saturation ratio (S i ) of a vapour,

and plotting the results on log-log axes, the slope gives the number of molecules of that

vapour in the critical cluster.

Besides laboratory measurements, the nucleation theorem has been applied in connection

of field measurements of atmospheric nucleation to predict the size of the critical

cluster (e.g. papers I–II). Just recently, there are new results that indicate that the

nucleation theorem may not hold in conditions in which there are local minima in the


Gibbs free energy surface (Vehkamäki et al., 2012). In that case, the conclusions on

the number of molecules should not be made based on the slope of log(J nuc ) vs. log(S i )

curve. So far, it is not known how the Gibbs free energy looks like in atmospheric nucleation.

However, a local minimum would be expected in atmospheric nucleation, as a

stable pre-nucleation cluster pool is observed to exist in field measurements (Kulmala

et al., 2007).

Also, Malila(2013)suggeststhatwhenapplyingthenucleationtheoremtonucleationin

thepresenceofbackgroundaerosol(asinatmosphericmeasurements), thecondensation

sink of the background aerosol may affect the interpretation of the results.

2.2.2 Atmospheric nucleation mechanisms

A wide range of experimental and theoretical evidence shows that sulphuric acid

(H 2 SO 4 ) is involved in atmospheric particle formation. In atmospheric aerosols, sulphate

is always observed, even if organic compounds often form the dominant part of

the aerosol mass (e.g. Jimenez et al., 2009). Sulphuric acid has low saturation vapour

pressure (< 10 −4 Pa, Ayers et al., 1980; Marti et al., 1997), which makes it eager to

nucleate and stay in the condensed phase in atmospheric conditions. Sulphuric acid

co-nucleates efficiently with water vapour, which is abundant in the atmosphere. Even

small amounts of H 2 SO 4 cause a huge increase of water nucleation rates in laboratory

experiments (Doyle, 1961).

On the basis of classical nucleation theory, two mechanisms for atmospheric nucleation

have been proposed: binary homogenous nucleation of sulphuric acid and water (Kulmala

et al., 1998a; Vehkamäki et al., 2002) and ternary nucleation of sulphuric acid,

water and ammonia (Korhonen et al., 1999; Napari et al., 2002a,b; Anttila et al., 2005;

Merikanto et al., 2007). Based on the acid-base interactions between the molecules (in

solution ammonia lowers the equilibrium vapour pressure of sulphuric acid), ternary

H 2 SO 4 -NH 3 -H 2 O nucleation happens easier (i.e. at lower saturation ratios) than binary

H 2 SO 4 -H 2 O nucleation; and both binary and ternary nucleation happen easier

than homogenous nucleation of H 2 SO 4 .

Binary H 2 SO 4 -H 2 O nucleation predicts nucleation rates well in the free troposphere

(Spracklen et al., 2005), but fails to expain nucleation in the boundary layer (e.g.

Spracklen et al., 2006; Chang et al., 2009b). However, lacking better nucleation theories,

binary H 2 SO 4 -H 2 O and ternary H 2 SO 4 -NH 3 -H 2 O nucleation theories have been

used widely to calculate nucleation rates in aerosol dynamical box models and global

models. In applying the theories, sometimes a correction factor has been used in order

to get the nucleation rates closer to the observed particle formation rates (Jung et al.,

2008). Ternary nucleation has been shown to work reasonably well in predicting the

occurence of nucleation events in polluted cities with high concentrations of sulphuric

acid and ammonia, but in terms of particle number it seems to produce too intense

nucleation events (Gaydos et al., 2005; Jung et al., 2008).

It has also been proposed that atmospheric nucleation could be purely kinetic i.e.


happen without any energy barrier (McMurry and Friedlander, 1979). In that case,

the nucleation rate would be determined only by the kinetic collision rate between

the molecules (with ∆G = 0 in Eq. 12), assuming that every collision results in the

formation of a stable cluster. For example, the rate of homogenous, barrierless kinetic

nucleation of sulphuric acid would equal the collision rate of H 2 SO 4 molecules. The

kinetic collision rate between molecules a and b is given by the equation (McMurry and

Friedlander, 1980; Seinfeld and Pandis, 2006):

K kin =

( 8kB T

πm ab

) 1/2π(ra

+r b ) 2 , (15)

where r a and r b are the radii of the reactant molecules and m ab = m a m b /(m a + m b )

is their reduced mass, T is temperature and k B the Boltzmann’s constant. The term

( 8k BT

πm ab

) 1/2 is the relative mean thermal velocity of the molecules and π(r a + r b ) 2 their

collision cross section. For H 2 SO 4 molecules at room temperature, the kinetic collision

rate is about 3 ·10 −10 cm 3 s −1 . The kinetic collision frequency sets the absolute

maximum for the possible nucleation rate in a system.

In atmospheric field measurements, a striking observation is that nucleation rates are

observed to correlate with sulphuric acid to the power between 1–2 (Weber et al., 1996;

Birmiliet al.,2000;Fiedler etal.,2005;Kulmalaet al.,2006, paper IandII).Thisisin

contradictionwithbinaryandternarynucleationtheories, whichpredict thecorrelation

exponents of > 10 and 5–10, respectively (Kulmala et al., 1998a; Vehkamäki et al.,

2002). In case of classical, homogenous nucleation with a simple form for the Gibbs

free energy curve, the correlation exponent corresponds to the number of molecules

in the critical cluster (Eq. 14, Vehkamäki et al., 2012). The slope of 2 was first

observed by Weber et al. (1996) and reported also by Birmili et al. (2000), but at that

time the observation did not recieve wider attention. Later, by H 2 SO 4 concentration

measurements at the Hyytiälä SMEAR II station, this connection was rediscovered

(Fiedler et al., 2005; Kulmala et al., 2006). In papers I–II this connection and its

implications were studied in detail.

The failure of binary and ternary nucleation theories, and the observation that particle

formation rates correlate simply with the first or second power of the sulphuric acid

concentration (J nuc ∝ [H 2 SO 4 ] n , with n = 1–2), has led scientists to develope empirical

parameterisations for the nucleation rate. These parameterisations, namely activation

and kinetic nucleation, were developed in papers I and II. They have been applied

quite widely in aerosol models (e.g. Spracklen et al., 2006) and been further developed

by Paasonen et al. (2009, 2010).

The correlation exponent (the slope of log(J nuc ) vs log([H 2 SO 4 ]) plot) has been taken

to represent the number of molecules in the critical cluster (see Eq. 14). Thus, atmospheric

observations have been interpreted so, that the critical cluster contains only

few (1–2) sulphuric acid molecules (papers I–II). However, this is too strong a conclusion

to make: as Vehkamäki et al. (2012) pointed out, if the Gibbs free energy curve

has a local minima at pre-nucleation sizes, the simple form of the nucleation theorem

(Eq. 14) is not valid. Thus, the correlation exponent n should not be interpreted


as the number of molecules of that kind in the critical cluster, but rather as giving

information about the rate limiting step in atmospheric nucleation. According to the

results of paper I and II, this rate limiting step is proportional to the sulphuric acid

concentration to the power 1–2.

Ion-induced nucleation, i.e. nucleation initiated by charged clusters, and ion-mediated

nucleation, including also the formationofneutral clusters by recombination of positive

and negative ions, have also been proposed as possible nucleation mechanisms in the

atmosphere (Yu and Turco, 2000). There is a constant charged cluster pool (of both

polarities) observed to exist in the atmosphere, and ion events similar to neutral new

particle formation events are observed (Kulmala et al., 2007; Hirsikko et al., 2011).

Theoretically, the presence of ions should enhance nucleation by introducing a local

minimum to the Gibbs free energy curve and by lowering its maximum. According to

current knowledge, ion-induced nucleation is expected to have only a minor contribution

to particle formation in the boundary layer, but possibly has some importance

in nucleation in the mid-troposphere (Hirsikko et al., 2011; Kirkby et al., 2011); although

some contradicting opinions also exist (Yu et al., 2008, 2010). In boreal forest

conditions, the contribution of ions to nucleation has been estimated to be about 1–10

% (Manninen et al., 2009; Gagné et al., 2010, 2012), while in other environments the

fraction has been observed to vary in the range 0.5–27 % (Manninen et al., 2010).

Incoastal areas, suchasMaceHeadinIreland, particleformationisobserved tohappen

through nucleation of iodine dioxide (OIO), emitted by algae when exposured to direct

sunlight (O’Dowd et al., 2002; O’Dowd and Hoffmann, 2005; Vuollekoski et al., 2009).

2.2.3 Laboratory measurements of atmospheric nucleation

In laboratory studies, performed mainly for the H 2 SO 4 -H 2 O system, it has been extremelydifficult

toobserve nucleationatconditionsmimicing theatmosphere(Viisanen

et al., 1997; Ball et al., 1999; Young et al., 2008; Benson et al., 2008; Brus et al., 2010).

Berndt et al. (2005) presented the first results of laboratory measurements, in which

nucleation was observed at close to atmosheric concentrations of sulphuric acid and

slopes approaching the atmospheric value of 2.

The comparison of various laboratory measurements and atmospheric measurements

(QUEST II, paper I) is presented in Figure 5 (Brus et al. 2010). Summarizing,

in the first decade of 2000’s, there were three discrepancies found in almost all laboratory

studies of H 2 SO 4 -H 2 O nucleation: i) The onset of nucleation requires much

higher concentrations than observed in the atmosphere. ii) The slope of the log(J)

vs log([H 2 SO 4 ]) plot is much higher than observed in the atmosphere, typically in the

range 4–8. iii) The results are sensitive to the production method of H 2 SO 4 : the nucleation

rates were much higher (or onset of nucleation happened at many orders of

magnitude lower concentrations) if H 2 SO 4 was produced in situ by reaction of SO 2 and

OH than if the H 2 SO 4 was evaporized from a liquid H 2 SO 4 sample. Especially the last

observation (iii) of high nucleation rates for in situ production of H 2 SO 4 as compared

to the liquid source was considered as a big mystery (Berndt et al., 2005; Brus et al.,


10 5 atmospheric (0 °C)

H 2

SO 4

from liquid samples

(20-30 °C)

10 3

J exp

[cm -3 s -1 ]

10 1

10 -1

10 -3

H 2

SO 4

from OH + SO 2


(20-30 °C)

10 -5

10 5 10 6 10 7 10 8 10 9 10 10 10 11

[H 2

SO 4

] [cm -3 ]

this work RH = 10,30,50%

Viisanen et al. RH = 38.2 and 52.3%

Wyslouzil et al. RH = 14 and 28%

Berndt et al. RH= 11,22,42,60%

Young et al. RH = 11,15,23%, 100 ppm

Young et al. RH = 15%, 1 ppm

Ball et al. RH = 2.3,4.7,7.5,15.3%,

Sihto et al. (2006) - atmospheric nucleation - Quest 2

Figure 5: Homogenous nucleation rate as a function of sulphuric acid concentration

for binary H 2 SO 4 -H 2 O system, obtained from different laboratory measurements and

comparison with the atmospheric data from the QUEST II campaign at the Hyytiälä,

SMEAR II station (paper I). Reprinted by permission from Brus et al. (2010).

2010). It led scientists to propose theories, that the nucleating agent would be some

other reaction product in the SO 2 +OH reaction pathway (see Sect. 2.3), such as HSO 3

or HSO 5 (Berndt et al., 2008; Laaksonen et al., 2008; Salonen et al., 2009).

In 2010, all three mysteries were solved at the same time by Sipilä et al. (2010). They

performed new laboratory measurements in a laminar flow tube applying novel measurement

technology for detection of small particles down to ∼ 1.3 nm: pulse-heightanalyzing

ultrafine-condensation particle counter (PHA-UCPC, Sipilä et al., 2009) and

particle size magnifier (PSM, Vanhanen et al., 2011). With these intstruments, nucleation

of H 2 SO 4 and H 2 O was observed at atmospheric conditions (onset of nucleation

at H 2 SO 4 concentration ∼ 10 6 cm −3 ) and with a slope 1.6–1.9.

The high onset H 2 SO 4 concentrations for nucleation and steep slopes of the log(J) vs

log([H 2 SO 4 ]) plot reported in earlier studies were explained to be caused by improper

instrumentation, which was not able to measure close to 3 nm sized particles with a

sufficient efficiency. The ultrafine condensation particle counter (UCPC, TSI 3025),

which was used in many studies, has a steeply rising counting efficiency at 3–6 nm.

This caused nucleation rates at different H 2 SO 4 concentrations (having also different


particle growth rates after nucleation) to be measured with different counting efficiencies,

resulting in an apparent increase of the nucleation rate with H 2 SO 4 concentration

and a high slope of the log(J) vs log([H 2 SO 4 ]) curve. Thus, in addition to the importance

of a suitable detector, also the growth rate (determined by H 2 SO 4 concentration

and residence time in the flow reactor) affects the results. Many earlier experiments

were performed with rather short residence times, resulting in a small growth rate, and

a large fraction of particles remaining so small that they were not counted.

The difference between liquid source and production of H 2 SO 4 from SO 2 +OH reaction

(”the sulphuric acid mystery”) was explained by different concentration profiles (as

a function of time) between these two cases: with a liquid, point-like instantaneous

source the concentration of H 2 SO 4 decreased steeply after nucleation, whereas in-situ

production yielded quite constant H 2 SO 4 concentration as long as the OH source (UVlight)

was on. In the former case, the growth rates were smaller than in the latter

case, resulting in a considerable fraction of nucleated particles not reaching a size big

enough to be detected.

In summary, proper instrumentation and high enough growth rates are required in

order to obtain correct results in nucleation experiments. Especially the Particle Size

Magnifier (PSM), which has close-to-unity counting efficiency for small particles, has

made it possible measure nucleation rates with good accuracy. It is possible that all

earlier laboratory experiments of nucleation are affected by the errors sources pointed

out by Sipilä et al. (2010), and therefore earlier laboratory results on sulphuric acidwater

nucleation should be interpreted with care.

Current thinking on atmospheric particle formation is as follows (Kulmala et al., 2013).

A more or less constant neutral cluster pool at 1–2 nm is observed to exist in the

atmosphere (Kulmala et al., 2007). Under certain, partly yet unidentified, conditions

these pre-critical clusters start to grow to bigger sizes. This happens normally during

daytime, with the participation of condensable vapours produced by photo-oxidation.

The most important vapour forinitiation ofthe growthof pre-nucleation clusters seems

to be sulphuric acid. Simultaneously with sulphuric acid, organic vapours start to

condense, possibly with the nano-Köhler mechanism, and speed up the particle growth


2.3 Formation and loss processes of sulphuric acid in the


Gas-phase sulphuric acid is produced in the atmosphere mainly through oxidation of

sulphur dioxide (SO 2 ) by OH radicals. The main sources of sulphur dioxide in today’s

atmosphere are emissions from fossil fuel burning and industry, global estimate for

emissions being 70 Tg(S)/year (Seinfeld and Pandis, 2006). Naturally SO 2 is emitted

from volcanic eruptions (global estimate 7–8 Tg(S)/year) and forest fires (global estimate

for emissions from biomass burning 2.8 Tg(S)/year) (Seinfeld and Pandis, 2006).

Over oceans, which lack extensive anthropogenic sulphur emissions, dimethylsulphide


(CH 3 SCH 3 , DMS) is the dominant source for SO 2 . DMS is produced by phytoplankton

and other marine organisms in the ocean and emitted into the atmosphere, where it

reacts with OH radicals and forms SO 2 (among other compounds). Altogether, DMS

is the largest natural contributor to the global sulphur flux into the atmosphere with

global emissions of 15–25 Tg(S)/year (Seinfeld and Pandis, 2006).

In the atmosphere, SO 2 reacts with hydroxyl radical (OH) in a sequence of reactions,

eventually forming sulphuric acid (e.g. Reiner and Arnold, 1994):

SO 2 +OH·+ M → HOSO 2 ·+ M (16)

HOSO 2 ·+ O 2 → HO 2 ·+ SO 3 (17)

SO 3 +H 2 O+ M → H 2 SO 4 + M. (18)

Here M represents a non-reactive molecule of the surrounding gas (in air typically N 2

or O 2 ), which is taking the excess energy released in the reaction. According to present

knowledge, this is the main mechanism for production of gas phase sulphuric acid in

the atmosphere. The second and third reaction (17 and 18) are fast, so that the rate

limiting step in thereaction scheme is reaction16. Thus the formationrateof sulphuric

acid is given by the rate of reaction 16: k 16 [SO 2 ][OH], where k 16 is the reaction rate


Hydroxyl radicals (OH·), in turn, are formed in photodissociation of ozone (O 3 )

molecules by ultraviolet (UV-B) radiation. The reaction mechanism is as follows:

O 3 +hν → O 2 +O,(λ < 1180 nm) (19)

→ O( 1 D)+O 2 ,(λ < 320 nm, UVB) (20)

O( 1 D)+H 2 O → 2 OH· , (21)

where O( 1 D) is the exited singlet state of the oxygen atom. This formation process

makes the OH concentration vary according to the intensity of UV-B radiation. The

formation depends also on the water vapour concentration ([H 2 O]), but the effect is

smalleraswatervapourisabundantinthetroposphere, andhaslessvariationcompared

to the solar UV radiation. OH is a highly reactive radical, with a lifetime of about 1

second, which reacts with almost all chemically reactive compounds in the atmosphere.

Despite the variety of loss processes for OH, the total OH concentration has been

observed to follow UV radiation with a good accuracy, both on diurnal and seasonal

time scales (Rohrer and Berresheim, 2006). Therefore, simply the intensity of UV

radiation can be used in modelling as a proxy for OH concentration, just scaled to a

proper maximum value (of the order 10 5 cm −3 –10 6 cm −3 ).

Owing to the high reactivity and short lifetime of OH radicals, measurement of OH

concentrations is rather difficult and requires indirect mass spectrometric techniques.


For modelling and data-analysis purposes, it is very useful that UV-B radiation can

be used as a proxy for OH. If measurements of UV radiation are not available, global

radiation (including all wavelengths) can be utilised in many cases, without losing too

much accuracy (Petäjä et al., 2009). However, the intensity of UV-B radiation on the

ground depends highly on the ozone column (the amount of O 3 in a vertical path),

implying that global radiation cannot be used as a proxy for UV-B in regions, where

the ozone column has strong seasonal variation.

Due to its low saturation vapour pressure, the main loss process for sulphuric acid

vapour is condensation onto pre-exisiting aerosol particle surfaces. In addition, nucleation

(if involving sulphuric acid) acts as a sink for sulphuric acid. Thus, the ambient

concentration of sulphuric acid vapour in the atmosphere (C v ) is governed by the equation:

dC v


= Q−CS ·C v −J nuc ·n ∗ sa , (22)

where Q is the source rate of H 2 SO 4 (the oxidation rate of SO 2 by OH: Q =

k 16 [SO 2 ][OH]), CS is the condensation sink (see Sect. 2.1.3), J nuc is the nucleation

rate and n ∗ sa is the number of H 2SO 4 molecules in the nucleating cluster.

As a loss process for H 2 SO 4 , the nucleation rate is of minor importance as compared to

condensation, even though nucleation is important with regard to particle formation.

Therefore as a first approximation the balance equation 22 can be rewritten:

dC v

dt =Q−CS ·C v (23)

=k 16 [SO 2 ][OH]−CS ·C v . (24)

Thus, the concentration of H 2 SO 4 is affected by the variations in SO 2 and OH concentrationaswell

as thecondensation sink of theaerosol particle population. As explained

above, OH varies according to sunlight (UV radiation). If the variations in SO 2 and

condensation sink are small compared to the variation of the daily cycle of OH, which

often is the case, then the H 2 SO 4 concentration varies approximately according to OH,

and thus according to sunlight.

Figure 6 shows an example of measured H 2 SO 4 concentration, together with the quantities

affecting the its formation: OH radical concentration, solar radiation (mainly

UV-B), SO 2 and water vapour concentration. The overall trend of the H 2 SO 4 concentration

follows roughly that of OH, but has some peaks that can be attributed to the

SO 2 emission peaks fromanthropogenic sources. Notethat Fig. 6 shows only thequantities

affecting the source rate and not the factors influencing the loss rate of H 2 SO 4 ;

the ambient H 2 SO 4 concentration is of course determined by the combined effect of

source and loss processes. Condensation on the background aerosol distribution is the

main loss mechanism for sulphuric acid, and abrupt changes in background aerosol

concentrations can cause rapid changes in the ambient H 2 SO 4 concentration.


Measuring the sulphuric acid concentration in the atmosphere is rather complicated,

and requires mass-spectrometric techniques (Berresheim et al., 2000, see Sect 3.1).

Therefore, based on the balance equation (24), proxies for estimating sulphuric acid

concentation from SO 2 concentration, UV-radiation and CS data have been developed

(Petäjä et al., 2009; Mikkonen et al., 2011). The proxy sulphuric acid concentration

was proven to agree reasonably well with measured H 2 SO 4 concentrations. However,

it should be noted that most of the proxies are constructed based on measurement

data from campaigns in spring and summer time, which may reduce the reliability

of the proxies in predicting the sulphuric acid concentrations during the cold season

(Mikkonen et al., 2011).

In clouds, aqueous reactions of SO 2 in cloud water (inside cloud droplets) is an important

source of sulphuric acid. Other sources for gas-phase sulphuric acid, in addition

to oxidation of SO 2 by OH, may also exist. Just recently, Mauldin III et al. (2012)

proposed a new, possibly important source of atmospheric H 2 SO 4 by reaction of SO 2

with Criegee Intermediates, highly reactive atmospheric biradicals (Welz et al., 2012).

7 x 106 Time (hours)

H 2

SO 4

concentration / Other scaled concentrations








H 2

SO 4

(cm −3 )

OH (cm −3 )

Glob x10 4 (W/m 2 )

UV−B x10 6 (W/m 2 )

SO 2

x10 7 (ppb)

H 2

O x10 6 (%o)

RH x 5x10 4 (%)

4 6 8 10 12 14 16 18 20 22 24

Figure 6: Example of the diurnal profile of sulphuric acid concentration (blue dots),

OH concentration, global and UV-B radiation, SO 2 concentration, water vapour concentration

and relative humidity (RH). The concentrations have been scaled to fit to

the same axis. The data were measured on March 25 th 2003 at the Hyytiälä SMEAR

II station.


2.4 Activation of aerosol particles to cloud droplets

Clouds are formed in the Earth’s atmosphere when water vapour in supersaturated

conditions (RH > 100 %) condenses on aerosol particles, thus forming cloud droplets.

In atmospheric temperatures, the homogenous nucleation of water vapour, i.e. formation

of water droplets without a seed aerosol particle, would require a relative humidity

on the order of 400 % (S ≈ 4). These conditions are never met for water vapour in the

Earth’s atmosphere, and therefore all cloud droplets are formed by heterogenous nucleation

and subsequent condensation of water vapour on top of a seed aerosol particle,

which happens at significantly lower saturation ratios. The supersaturated (supersaturation

is defined as SS = S − 1 and expressed typically in %) conditions required

for cloud formation (SS > 0 %) are typically met in the upper boundary layer and in

the free troposphere, where air temperature is low enough to cause a low saturation

vapour pressure and a high enough water saturation ratio. The seed aerosol particles

are called cloud condensation nuclei, abbreviated as CCN.

Activation of an aerosol particle means that water starts to condense irreversibly on

the particle, making it grow into a cloud droplet with size on the order of µm. The

critical diameter is the smallest particle size that gets activated at a certain water

supersaturation. Vice versa, the critical saturation ratio is the saturation required for

a certain particle size and composition in order to start irreversible water condensation

(i.e. activate the particle). The equilibrium between the aerosol particle and water

vapour is described with the Köhler equation:

( 4Mw σ


S eq = a w exp , (25)

RTρd wet

where S eq is the water equilibrium saturation ratio, a w is the water activity of the

solution (a w = γ w x w , in which γ w is the activity coefficient and x w is the mole fraction

of water in the solution), M w is the molar mass of water, σ is the surface tension of

the droplet, R is the universal gas constant, T is temperature, ρ is the density of the

solution and d wet is the wet particle diameter. Typically, for the surface tension and

the density the values for water are used (dilute solution).

The Köhler equation combines the curvature effect (the Kelvin effect, Eq. 1), which

increases the equilibium vapour pressure of water on top of the droplet surface, and

the solute effect (the Raoult effect), which decreases the equilibrium vapour pressure

above a solution. As a result, a curve with one maximum is obtained. The Köhler

curve is calculated for a certain dry aerosol particle. The critical saturation ratio S crit

(with a certain dry particle size and composition) corresponds to the maximum of the

Köhler curve: at saturation ratios bigger than S crit , water condenses continuously onto

the particle. At a certain saturation ratio, the corresponding dry particle size can be

obtained by iteratively solving the Köhler-equation: this particle size is the smallest

dry particle that gets activated at the given saturation ratio. This particle size is called

the critical diameter or threshold diameter for cloud droplet activation (Kerminen et

al., 2012).


The critical diameter for cloud droplet activationdepends on the water saturation ratio

and on the chemical composition of the aerosol particle. In the atmosphere, the critical

diameters are typically in the range 50–100 nm.

Particle hygroscopicity means its ability to take up water in subsaturated conditions

(S < 1 of RH < 100 %) and it is described by the hygroscopic growth factor:

g a = d wet

d dry

, (26)

where d wet and d dry are diameters of the wet (at certain RH) and dry particle, respectively.

Particle hygroscopic growth factors are measured by Hygroscopic Tandem

Differential Mobility Analyser (HTDMA) for a specific relative humidity (RH).

Another formulation of the Köhler equation was presented by Petters and Kreidenweis

(2007), who expressed the thermodynamic properties of the particle (the solute

effect described by water activity a w ) with the aid of a hygroscopicity parameter κ,

determined by:


a w

= 1+κ V dry

V water

, (27)

where V dry is the volume of dry particulate matter and V water is the volume of water in

the solution (V wet = V dry + V water ). With this parametrisation for the water activity,

and a couple of other assumptions, the kappa-Köhler equation can be derived:

S eq (d wet ) =

d 3 wet −d3 (


d 3 wet −d 3 dry (1−κ) exp 4Mw σ



, (28)

RTρ w d wet

where ρ w is the density of water and σ w is the surface tension of water. Note that particle

density has disappeared from the equation and all the physico-chemical properties

(except the diameter) of the aerosol are captured in the κ-parameter.

The Köhler equation is typically applied at saturation ratios S > 1 necessary for

cloud formation, even though it is valid also for S < 1. The idea of kappa-Köhler

equation is to apply it also for S < 1, thus covering the full range of saturation ratios

from hygroscopic growth (S < 1) to cloud droplet activation (S > 1) (Petters et al.,

2009). By expressing the wet diameter with the aid of the hygroscopic growth factor

(g a = d wet /d dry ), the κ-Köhler equation becomes:

S eq (d dry ) =

g 3 a −1

g 3 a −(1−κ) exp ( A

g a d dry


,A = 4M wσ w

RTρ w

. (29)

Using HTDMA data on particle hygroscopic growth factors at subsaturated conditions

with a known saturation ratio, the κ-parameter of the aerosol can be determined by


solving it from the kappa-Köhler equation. The higher the value of κ, the more hygroscopic

the aerosol is. Kappa-parametershave been determined in laboratoryconditions

for pure organic aerosol (κ org = 0.1, from α- and β-pinene) and for ammonium sulfate

(κ as = 0.6) as well as in field measurements at several locations (e.g. Gunthe et al.,

2009; Dusek et al., 2010). For purely non-hygroscopic aerosol, such as mineral dust, κ

=0.01–0.08(Koehleretal.,2009). Estimates fortheglobalmeanofκare0.27±0.21for

continental and 0.72±0.24 for marine aerosol (Pringle et al., 2010). The kappa-values

can then be applied to predict the CCN activity of the studied aerosol: to estimate the

critical diameter for cloud droplet activation using the kappa-Köhler equation.


3 Methods

This thesis combines analysis of field measurement data and aerosol dynamical simulations.

This section describes the experimental data, the main data analysis methods

and the aerosol dynamical model used in this study.

3.1 Experimental data


Finland (papers I-III, VI). In paper II also measurement data from the Heidelberg

station in Germany were analyzed.

TheSMEARII(StationforMeasuringforestEcosystem-Atmosphere Relations)station

is located at the Hyytiälä Forestry Field station, about 60 km north-east of the city

of Tampere ( The environment represents a

typical Finnish rural area, with fields and large areas of mixed forest, dominated by

coniferous trees (spruce and pine). The station is affected also by pollution from a

nearby city (Tampere) and industrial sites. The measurement station itself is inside a

roughly 40-year old Scots pine (Pinus Sylvestris L.) forest. The leading idea in setting

uptheHyytiälämeasurement stationhasbeentomeasure”everythingpossible”related

to forest and atmosphere: from soil moisture to photosynthetic gas exhange between

vegetation and atmosphere, pollutant trace gases as well as aerosol particle and ion

concentrations. The Hyytiälä SMEAR II station is a unique site for multi-disciplinary

research with high synergetic effects arizing from the possilibity to combine different

sources of data (Hari and Kulmala, 2005).

Particle size distribution measurements

Particle size distributions were measured by a Differential Mobility Particle Sizer

(DMPS) setup. Figure 7 presents the setup operating at the Hyytiälä SMEAR II

station (Aalto et al., 2001) at the present configuration; some details (for example the

charger and the inlet) have changed since the measurements in 2003 reported in paper


The inlet is situated at a height of about 8 m from the ground (until 21.9.2004 the

inlet was at 2 m height) and the measurement devices are located inside a cottage. The

air sample is taken using a TSP-inlet (Total Suspended Particles). Particle sizing is

performed using differential mobility analysers (DMA), which classify aerosol particles

according to their electrical mobility. The sheath flow for the DMA is dried, so that

inside the DMA aerosol is ”dry” at a relative humidity < 30 %. For sizing with DMA

the particles must be charged. This is achieved using a charger with a 14 C betasource

(until 14.10.2008 the charger was 85 Kr), in which the aerosol achieves a charge

equilibrium (for this reason the charger is also called a ”neutralizer”). At a certain

voltage between the plates of a cylindrical DMA, a certain particle size d p ± ∆d p is





C ( )








sheath flow


closed loop)







d p

= 3...50 nm


TSI 3025


TSI 3010

d p

= 10...1000 nm




* Inversion algorithm:

- d p

- charging probability

- sampling losses

- CPC detection efficiency

Figure 7: A schematic picture of the DMPS (Differential Mobility Particle Sizer) setup

at the Hyytiälä SMEAR II station.

passing through the DMA and this concentration is measured with a condensation

particle counter (CPC). By scanning the voltage of the DMA as a step function, a

size spectrum is obtained. To cover the particle size range 3 nm–1 µm, two parallel

DMA+CPC systems are needed (twin-DMPS): the first one measures the range 3–

50 nm and the second the range 10 nm–1 µm (until 8.12.2004 the upper limit was

500 nm). The voltage scanning is performed in different directions in DMA1 (small

particles) and DMA2 (bigger particles) in order to measure the overlapping part of the

spectra temporally as close as possible. One scan of the whole size distribution takes

about 10 min., defining the time resolution of the twin-DMPS system.

The raw output data from the DMPS is the mobility distribution of the aerosol particles,

which were negatively charged after the 14 C-charger. By applying an inversion

algorithm, the electrical mobilities are converted to particle sizes and the concentration

of neutral aerosol particles (corresponding to the ambient situation before the charger)

is deduced using the particle charging probabilities. The inversion algorithm takes also

into account the estimated particle losses happening in the sampling lines.

The Hyytiälä DMPS system measures particle size distribution with 10 min. time

intervals in 38 channels logarithmically distributed between 3 nm and 1 µm (diameter)

(previously 23 size channels between 3 and 500 nm). The constant RH (about 30 %)


allows for comparison of size distributions despite of big variations in ambient RH.

Sulphuric acid concentration measurements

The gaseous sulphuric acid concentrations were measured with a Chemical Ionisation

MassSpectrometer(CIMS)(Hankeetal.,2002;Fiedleretal.,2005). Themeasurements

during the QUEST II–IV campaigns were performed by a group from Max Planck

Institute (MPI-K Heidelberg) led by Frank Arnold.

The measurement principle of CIMS is to convert sulphuric acid molecules (H 2 SO 4 )

to ions by a chemical reaction, after which the concentration of the product ions can

be measured by a mass spectrometer (Fiedler et al., 2005; Aufmhoff et al., 2011).

The instrument consists of an ion source (NO − 3 (HNO 3) n -ions), a flow reactor and the

quadrupole mass spectrometer, as well as a H 2 SO 4 source for the calibration of the

instrument. In the flow reactor, a fast ion-molecule reaction between NO − 3(HNO 3 ) n -

ion and H 2 SO 4 happens:

NO − 3 (HNO 3 ) n +H 2 SO 4 → HSO − 3(HNO 3 ) m +(n−m)HNO 3 .

The concentrations of reagent (NO − 3(HNO 3 ) n ) and product ions (HSO − 3(HNO 3 ) m )

are measured with the quadrupole MS. The concentration of H 2 SO 4 can then be inferred

from the ratio of these concentrations when the reaction rate constant is known

(Berresheim et al., 2000). The detection limit for sulphuric acid was 1·10 5 cm −3 and

the relative measurement error 30 %. The time resolution of the spectrometer was less

than 1 s, but the data was averaged over 60 s in order to reduce statistical error.

The CIMS instrument has also been used to measure other atmospheric trace gases,

such as volatile organic compounds (Sellegri et al., 2005) and OH (Berresheim et al.,

2000; Petäjä et al., 2009; Aufmhoff et al., 2011).

Cloud condensation nuclei and hygroscopicity measurements

The cloud condensation nuclei concentrations were measured at ground level at the

SMEAR II station by a CCN-counter (CCNC, model DOC-0086 by Droplet Measurement

Technologies). The CCN-counter mimics the conditions inside a cloud and measures

the number of particles that are activated for cloud droplets at a certain water

supersaturation. Bycomparingthenumberofactivatedparticles(the”clouddroplets”)

to the total particle concentration, the activated fraction can be determined; by comparing

the number of activated particles with the particle size distribution, an estimate

of the critical diameter for cloud droplet activation can be obtained. The instrument

operates at supersaturations from 0.07 % to 3 %, thus capturing the typical conditions

prevailing in cloud formation. At the SMEAR II station the CCNC instrument

operates at water supersaturations of 0.1–1.0 %.


sample flow

sheath flow

wetted walls





d T/dz

( T > T > T )

3 2 1

T 1

T 2




T p

( T )

A, sat A

T , p ( T )


diffusion of heat

and water vapour










( TB



( T )


T 3


Figure 8: A schematic picture of the Cloud Condensation Nucleus Counter (CCNC).

The operation of the CCN-counter is based on the principle that diffusion of heat in

air is slower than diffusion of water vapor (Roberts and Nenes, 2005; CCNC manual).

Figure 8 presents a schematic of the CCN-counter. An aerosol sample is led through a

cylindrical flow chamber, which has wetted walls for providing a constant water vapour

source and a constant, increasing temperature gradient over the cylinder height. The

water vapour(saturationconcentration prevailing above thewetted walls) andheat diffusefromtheoutershell

inwards, tothecenter ofthecylinder. Because watermolecules

diffuse more quickly than heat (due tothe smaller size ofaH 2 O molecule in comparison

to air molecules N 2 and O 2 ), the vapour concentration and heat (temperature) at the

centerline (point C) originate from different locations on the wall (points A and B): at

point C there is a concentration p sat (T B ) of water vapour at a temperature of T A . Because

T A < T B , the vapour is supersaturated at C: S = p sat (T B )/p sat (T A ) > 1. Due to

the constant temperature gradient, in the centerline of the cylinder there is a constant

supersaturation prevaling (actually S > 1 almost everywhere inside the cylinder).

The supersaturation at the centerline of the flow tube, where the aerosol sample is

flowing, depends on the temperature gradient between bottom and top of the cylinder,

flow rate and pressure. While the flow rate and pressure are kept constants, the

operating supersaturation is chosen by adjusting the temperatures T 1 , T 2 and T 3 . At

a given supersaturation, particles bigger than the critical diameter activate (i.e. water

starts to condense on them irreversibly) and continue to grow fast as they travel in

the centerline of the flow tube (to about to some µm in diameter). The concentration


of activated cloud droplets is then counted by an optical particle counter (OPC). The

OPC detects particles of size 0.75–10 µm and allows also a rough sizing of the droplets.

In this work, the droplet size detection possibility of the CCNC was not utilized.


Mobility Analyser (HTDMA) (Ehn et al., 2007). The principle of the instrument is

to first select a dry aerosol particle size with a DMA, then lead the sample through a

humidifier with a specified RH (RH < 100 %), and after that measure the humidified

size distribution. From the humidified distribution (which may be uni- or multimodal

depending onthe mixing state of the aerosol) the growth factor distribution is obtained

by dividing it by the original dry diameter (see Eq. 26).

Meteorological and gas data

In addition to the data described above, accompanying data of meteororological

variables (T, p, RH) and common trace gases (H 2 O, SO 2 etc.) were utilized

in the data analysis (for the measurement methods see SMEAR webpage

3.2 The calculation of particle formation rate

The particle formation rate J dp is defined as the number of particles of diameter d p

formed per cm 3 per second. In atmospheric particle formation studies, the common

terminology is as follows:

- ”the nucleation rate” is the formation rate of the smallest stable clusters. According

to current knowledge, the size of the nucleated clusters is 1–2 nm (Kulmala et al.,

2007). The nucleation rate is marked as J nuc , J 1 or J 2 (subscript referring to the size

in nm). The unit is particles/(cm 3 s) = cm −3 s −1 .

- ”the particle formation rate” refers to the formation rate of particles at some other

size than the nucleation size. Often the size of interest is set by the lower detection

limit of the measurement instrument, e.g. 3 nm (in diameter) for the conventional

aerosol instruments (CPC, DMPS and SMPS). The rate is marked as J dp , for example

J 3 , and the unit is cm −3 s −1 . Sometimes also the term ”apparent particle formation

rate” is used.

This distinction of the nucleation rate and the particle formation rate was first suggested

by McMurry and Friedlander (1979) and has become a common convention in

aerosol science.


3.2.1 Particle formation rate at 3 nm

Let us consider first a continuous particle size distribution at the diameter d p = 3 nm.

We can write for the number concentration at 3 nm:



∣ = dN ∣ ∣∣3 dd p

∣ = n(d p ) ∣ 3 dd p dt 3 3

×GR 3 ≡ J 3 , (30)

where n(d p ) ∣ ∣


= dN/dd p

∣3 is the size distribution function at d p = 3 nm and GR 3 =

dd p /dt is the diameter growth rate of 3 nm sized particles. This is the mathematical

definition of the particle formation rate at 3 nm, J 3 (provided that self-coagulation is

not significant). In words, J 3 means the flux of particle concentration on the diameter

axis at size d p = 3 nm (see Fig. 9). In principle, if we know the size distribution and

growth rate accurately at 3 nm, one could apply directly Eq. 30 to calculate J 3 , or in

practice with an approximated expression (Weber et al., 1996):

J 3 = n(d p ) ∣ ∣


×GR 3

∼ =


∆d p

∣ ∣∣3

GR 3 , (31)

where ∆N is the particle number concentration at a narrow size range ∆d p around

d p = 3 nm. This expression has been used by some researchers to estimate J 3 from

PHA (Pulse Height Analysis method) and nano-SMPS (Scanning Mobility Particle

Sizer, similar to DMPS) data, with particle size ranges ∆d p = 3–4 nm (PHA) and ∆d p

= 3–6 nm (nano-SMPS) (Weber et al., 1997; Kuang et al., 2008).

We adopt a sligtly different method for the calculation of the particle formation rate,

in which the particle size distribution is explicitly considered as a discrete distribution

(Kulmala et al., 2001a). This corresponds to the real case of aerosol measurements,

where the particle size distribution is measured as discrete channels. For one size bin

at diameter d p,i we can write the balance equation for the number concentration N i :

dN i


= J i −J i+1 −Coag. loss, (32)

where J i is the flux into the size bin (= formation rate of particles at the lower limit of

the bin), J i+1 is the flux out of the size bin at the upper limit of the bin, and Coag. loss

is the loss rate of particles due to coagulation with other (larger) particles of the size

distribution. The quantities are depicted in Figure 9. This balance equation can be

derived from the general dynamic equation (Eq. 4) by integrating it over the particle

diameter range ∆d p,i (paper V).

Applying Eq. 32 for the particle size range 3–6 nm and using the definition of particle

formation rate (J dp , Eq. 30) we get:


Figure 9: A schematic picture of the quantities affecting the particle concentration N i

in a size bin d i ...d i+1 .

dN 3−6

= J 3 −J 6 −Coag. loss



≈ n 3 GR 3 −n 6 GR 6 −CoagS dp=4nmN 3−6 , (34)

where the coagulation loss has been expressed with the aid of the coagulation sink.

To simplify the calculation, the coagulation loss for 3–6 nm particles is approximated

by the coagulation loss of 4 nm sized particles having the concentration N 3−6 (4 nm

is close to the geometric mean of 3 and 6 nm). The size range 3–6 nm was chosen,

because it is small enough to be considered as freshly nucleated, but large enough to

achieve relatively good statistics for the number concentration. In addition, the same

size range has been used in earlier studies of atmospheric nucleation (Weber et al.,


By solving Eq. 34 for J 3 we obtain:

J 3 = dN 3−6


+n 6 GR 6 +CoagS dp=4nmN 3−6 . (35)

Calculating the size distribution function from N 3−6 as n 6 = ∆N 3−6 /∆d p , using the

growth rate determined from the DMPS data (for size range 3–10 nm) and approximating

differentials with finite differences, we get:

J 3 = ∆N 3−6



N 3−6

(6−3) nm GR DMPS +CoagS dp=4nmN 3−6 . (36)

This is the equation which is used to calculate the formation rate of 3 nm particles

from the size distribution data measured with DMPS.


3.2.2 Estimation of the nucleation rate from the apparent particle formation


After nucleation, the newly formed particles grow by condensation and are scavenged

by coagulation with the pre-existing particle size distribution. In the absence of coagulation

(and other removal mechanisms) J 3 would equal the nucleation rate J nuc

(J 3 (t + ∆t) = J nuc (t)), after a time delay ∆t associated with the growth time from

nucleated size (d nuc ) to 3 nm. With coagulation scavenging, which acts as a sink for

particles during the growth from d nuc to 3 nm, the particle formation rate at some

other diameter is always smaller than the real nucleation rate: J 3 (t+∆t) < J nuc (t).

Inmeasurements, thetypical lower detectionlimit (e.g. DMPSandSMPSinstruments)

for particle size is d p = 3–10 nm, even though recent advances in aerosol instrumentation

have pushed the detection limit down to 1–2 nm (NAIS, PSM; Asmi et al., 2009;

Sipilä et al., 2009; Vanhanen et al., 2011). Anyhow, a vast majority of aerosol measurements

are still done with a lower detection limit of 3 nm. For nucleation studies, it

would be useful to be able to estimate the real nucleation rate at 1 nm or 1.5 nm from

the observed apparent particle formation rate at some greater diameter, say 3 nm.

The original Kerminen-Kulmala formulation

Kerminen and Kulmala (2002) presented a formula connecting the apparent particle

formation rate J dp at a diameter d p with the nucleation rate J nuc at diameter d nuc :


J dp (t+∆t) = J nuc (t)exp

(γ CS′ 1

− 1 ))

, (37)

GR d p d nuc

where GR is the particle growth rate (in units nm/h)) and nucleated size d nuc and

particle size d p are expressed in meters (m). CS ′ (units m −2 ) is directy proportional

to the condensation sink (CS):

CS ′ = 1 2


β m,i d p,i N i = CS

4πD v

, (38)

where D v is the diffusion coefficient of the condensing vapour (assumed to be sulphuric

acid). γ in Eq. 37 is a fitting parameter with an approximate value of 0.23 nm 2 m 2

h −1 .

Equation 37 takes into account the competition between coagulational scavenging

(through the term CS ′ ) and condensational growth (through the term GR) during

growth from the nucleated size d nuc to a size d p . The exponential represents the probability

that a nucleated cluster of size d nuc , subject to coagulation and condensational

growth, will survive to the size d p . The time delay ∆t arises from the the growth time

from d nuc to d p : ∆t = (d p −d nuc )/GR.


The Kerminen-Kulmala equation (referred to as the K-K equation) was derived under

three main assumptions (Kerminen and Kulmala, 2002):

(i) coagulation to background aerosol is the only sink for nucleated particles, i.e.

self-coagulation and dry deposition are neglected. Neglecting self-coagulation is

justified if concentations are below 10 5 –10 6 cm −3 .

(ii) the particle growth rate is constant during growth from d nuc to d p .

(iii) background aerosol (i.e. CS’) stays constant during the growth from d nuc to d p .

The K-K equation was derived analytically, but it contains a fitting parameter γ,

which arises from expressing the coagulational scavenging (the coagulation sink) in

terms of the condensation sink. The reason behind this is that at small particle sizes,

approaching the molecular size, coagulation can be thought of as condensation of nmsized

particles onto the background distribution. Also, CS is often the quantity that

is calculated from measured particle size distributions, and is more straightforvard to

calculate than CoagS.

The Brownian coagulation coefficient of d p -sized particles is proportional to d −κ

p , where

the exponent κ is in the range 1.5–2 (Kerminen and Kulmala, 2002; Seinfeld and

Pandis, 2006). Kerminen and Kulmala assumed κ = 2, and thus the coagulation sink

of d p -sized particles can be expressed as:

CoagS dp = γ ′ ·CS ′ (


d p

) 2

= γ ·CS ′ ( 1

d p

) 2

. (39)

The proportionality factors γ ′ and γ connect CoagS(d p ) to CS ′ . Based on fittings to

data from aerosol dynamical simulations, Kerminen and Kulmala (2002) presented a

parametrisation for γ which depends weakly on many factors such as temperature and

nucleated particle density. However, usually an approximate value of 0.23 nm 2 m 2 h −1

is accurate enough for atmospheric particle formation studies.

In connection with aerosol measurements, the K-K equation is often used in the reverse

direction, i.e. to estimate the nucleation rate fromthe observed particle formationrate:


J nuc (t) = J dp (t+∆t)exp

(−γ CS′ 1

− 1 ))

. (40)

GR d p d nuc

Usually the equation is used at the lower detection limit of particle size distribution

measurements, i.e. at d p = 3 nm, and with nucleation size in the range 1–2 nm.

The Kerminen-Kulmala equation has proved to be very useful in testing and developing

nucleation theories (e.g. papers I and II). With this equation we can obtain

quite reliable estimates of the actual nucleation rates, when accurate measurements of


nucleation rate at the real nucleation size (1–2 nm) are missing. Another important

use of the K-K equation is in large scale atmospheric models, where it (in the form of

Eq. 37) can be used to transfer nucleation rates to particle formation rates e.g. at 3

or 10 nm, without the need to model all the initial steps of particle growth in detail.

The Kerminen-Kulmala equation corresponds to a similar formula presented by Mc-

Murry and Friedlander (1979). However, the Kerminen-Kulmala equation has become

more popular, most probably because it is easier to apply and uses two main quantities

that are determined in new particle formation event studies, namely GR and CS. In

their paper, McMurry et al. (2005) present a rigorous examination on the connections

between these two similar methods to estimate nucleation rates from the apparent

particle formation rate.

The revised form of the Kerminen-Kulmala equation

Afterwards, Lehtinen et al. (2007) have presented a revision for the Kerminen-Kulmala

equation. In their formulation, two improvements were made: the coagulation scavenging

is calculated explicitly from the coagulation sink (not through the condensation

sink) and a more accurate expression for coagulation sink is used. Instead of assuming

the coagulation sink to be proportional to the square of the particle diameter (Eq. 39),

the exponent is kept as a free parameter:

CoagS(d p ) = CoagS(d nuc )·



d nuc

) m

, (41)

where CoagS(d nuc ) is the coagulation sink of nucleated particles and m = −κ. By

solving this, an equation for the exponent m is obtained:

m = log[CoagS(d p)/CoagS(d nuc )]

. (42)

log[d p /d nuc ]

By this equation the value of exponent m (typically in the range [−2,−1.5]) can be

calculated directly from the particle size distributions. At the Hyytiälä SMEAR II

station, m varies in the range [−1.75,−1.5] with a mean value −1.7 (Lehtinen et al.,


The main improvement of this approach is that we get rid of the fitting parameter

γ, which in the K-K formula was adjusted based on modelling results. With the new

expression for the CoagS (Eq. 41), a new equation for the apparent particle formation

rate was obtained:


( )

CoagS(d nuc )

J dp (t+∆t) = J dnuc (t)exp −γ d nuc ,


where γ =

[ ( ) ] m+1

1 dp

−1 .(43)

m+1 d nuc

Note that here the parameter γ is a dimensionless parameter, and different from the

one in the original form of the K-K-equation. To get the units correct in this equation,

the CoagS has to be expressed in units h −1 , if the GR is in nm/h.

There are a couple of advantages in the revised version of the K-K equation. First, it

is conceptually more clear to use directly the coagulation sink instead of condensation

sink for calculating the coagulational scavenging. Also, the dependence of the condensation

sink on the diffusion properties of the condensing vapour is avoided. Second,

the coagulation scavenging is calculated more accurately. Third, the varying ambient

conditions are more easily taken into account, as the parameter m can be calculated

explicitly from the particle size distributions.

The revised version of the K-K equation (Eq. 43) is applied as follows. First, the

proper value of m is calculated from experimental particle size distribution data, to

represent the conditions of the studied case. The parameter m can be determined for

theaverage conditions of the station, or even separately for each new particle formation

event. Then the particle formation rate is evaluated using Eq. 43.

In addition to the modification by Lehtinen et al. (2007), a few other improvements

have been presented to the original K-K equation. Kerminen et al. (2004) presented a

formulation which accounts for, in addition to sulphuric acid, an organic vapour contribution

to the particle growth, and allows for a time-dependent growth rate. Anttila

et al. (2010) included in their parameterisation the effects of self-coagulation of freshly

nucleated particles. However, in most cases of atmospheric particle formation (when

the nucleation rates are not too high and the growth rate is not varying fast), the original

formulation by Kerminen and Kulmala (2002) or Lehtinen et al. (2007) is accurate


In the articles included in this thesis, the original formulation of the K-K equation

(Eqs. 37 and 40) was used, except in paper III, where the new, revised formula by

Lehtinen et al. (2007, Eq. 43) was applied. In future studies, the use of the revised

formula is recommended.

3.3 Evaluation of the calculation method of J 3

Equation 36 has been widely used for estimating the new particle formation rate from

measured particle size distribution data. In paper V we studied the accuracy of this

calculation method to predict the actual particle formation rate at 3 nm. The study


wasbasedonsimulationsmadewiththeUniversity ofHelsinki Multicomponent Aerosol

model (UHMA; see Section 3.4 for description of the model). Using simulated data,

the exact particle formation rate at 3 nm (J 3 ) as well as other model conditions are

known, and can be compared with the estimate given by approximate equation (36).

For the evaluation study, a new particle formation event, similar to the observed new

particle formation events, was produced with the UHMA model (Fig. 10). The nucleation

mechanism was activation nucleation (see Sect. 4.2) at a diameter of 1.5 nm,

and particle growth was caused by sulphuric acid and a non-volatile organic vapour.

Both vapours had a sinusoidal-like diurnal profile, while the concentration of organic

vapour was significantly higher than that of H 2 SO 4 , so that particle growth was caused

mostly by the organic vapour. The background particle distribution corresponded to

a typical case in Hyytiälä.

From the simulated data, the particle formation rate at 3 nm was calculated by Eq.

31 (corresponding to the Eq. 3 in paper V):

J 3 = ∆N

∆d p

∣ ∣∣3

GR 3 , (44)

where ∆N/∆d p | 3 is the size distribution function and GR 3 is the growth rate at d p

= 3 nm. Let us denote this formation rate as ”J 3 , exact”. This formation rate is

called ”exact”, since we get the values of particle size distribution ∆N/∆d p and the

growth rate GR 3 directly from the simulated data, and the formula (Eq. 31 or 44) does

not contain any approximations despite the calculation of n 3 from the discrete model

data. In the simulated case we used 60 size sections (as compared to 23 size sections

in the measured data), so discretization will have only a minor effect on the results.

Consequently, J 3 calculated by Eq. 44 can be considered to represent the actual J 3

quite accurately.

For evaluation of the validity of the J 3 calculation methods, Eq. 36 was applied to

the modelled event in the same way as for a measured particle formation event. The

∆N 3−6 /∆t, N 3−6 and CoagS dp=4nm were obtained directly from the simulated discrete

particle size distribution. The growth rate was calculated in two ways: i) directly from

the condensation rate of vapours, calculated as an average for particle sizes 3–7 nm, or

ii) by estimating a constant GR from the time evolution of the nucleation mode peak

diameter between 3–7 nm (GR = ∆d p /∆t). The first method (i) is applicable only

withthesimulated dataandgivesatime-dependent (but size-averaged) GR. Thelatter

method (ii, denoted as ”constant GR”) is the way the growth rate is determined from

measured new particle size distribution data (Hirsikko et al., 2005). Thus Eq. 36 with

the ”constant GR”-growth rate corresponds to the measurement case for calculating

J 3 .

The investigation showed that Eq. 36 (corresponds to Eq. 5 in paper V) gives a fairly

good estimate for new particle formation rate J 3 (see Fig. 10). The equation tends to

overestimate the particle formation rate J 3 , as compared to the exact values given by

Eq. 44, with an error of 10–20 %. Surprisingly, the J 3 estimation is better with the


Formation rate [cm −3 s −1 ]









Eq. 3

Eq. 5

Eq. 5 (constant GR)

Eq. 6




6 8 10 12 14 16 18

Time [h]

Figure 10: Simulated new particle formation event (left) and comparison of different

methods to calculate the particle formation rate J 3 during this event (right) (same as

Figs. 2 and 3 in paper V). Note that the legend refers to the equations of paper V.

”constant GR” growth rate, i.e the method used in analyses of measured data, than

with the more exact, time-dependent GR calculated from the simulation. This effect

is most probably caused by a fortuituos error cancellation.

The sensitivity studies presented in paper V indicated that most of the error in J 3

calculation (Eq. 36) can be attributed to coagulation and its various effects. Between 3

and6nmparticlesexperience significant coagulationscavenging tobackground aerosol,

with the coagulation rate decreasing as particles grow larger (from 3 to 6 nm). The

coagulational scavenging term CoagS 4 × N 3−6 is only a rough approximation to the

coagulation rate of 3–6 nm particles. The other main error source is the calculation of

J 6 : J 6 = n 6 GR 3−7 (Eq. 31), where the size distribution function was approximated as

n 6 = N 3−6 /(6−3)nm. Due to decreasing coagulation rates, more particles at 3 nm are

removed by coagulation than at 6 nm, and thus the mean number concentration in the

size range 3–6 nm, N 3−6 , is not the best estimate for n 6 close to upper diameter 6 nm.

A better choice would be to estimate n 6 from a size range closer to 6 nm, e.g. 5–7 nm:

n 6 ≈ N 5−7 /(7−5)nm.

Applying themoreaccuratecalculation forn 6 improved theestimate ofJ 3 significantly;

the error in J 3 was reduced to only 6 %.

In this study, the effect of self-coagulation was neglegted, but the effect should be small

at these concentrations (significant only at very high number concentrations).

With this modification to the n 6 estimation, a new formula for calculating J 3 was

proposed (corresponds to Eq. 6 in paper V):

J 3 = ∆N 3−6


+ N 5−7

(7−5)nm GR 3−7 +CoagS dp=4nmN 3−6 . (45)

Because this improvement is straightforward to implement, in paper V we recommend

this new form of J 3 equation to be used in the analyses of experimental particle

formation data.


It is clear that also the growth rate GR 3−7 , calculated from the time evolution of the

nucleation mode, has inaccuracies (Leppä et al., 2011), but those errors are hard to

quantify andso far no other reasonable way to estimate growth rates exists. Simulation

results showed that the J 3 calculation was quite sensitive to the value of the growth


This study (paper V) was to our knowledge the first attempt to estimate the validity

of the particle formation rate calculations, which have been perfomed for a wide range

of measurement data from different locations. It gave confidence that the method

used to estimate J 3 gives generally fairly good results, and the magnitude of error is

acceptable. However, the study was based on only one type of new particle formation

event, although several sensitivity tests on the simulation conditions were performed.

Tosettheresultsonamoresolidbase, amoredetailedstudywithavarietyofsimulated

new particle formation events would be needed in order to find out if these results

are statistically valid, and what is the uncertainty in the calculation of J 3 for larger


3.4 University of Helsinki Multicomponent Aerosol model

Aerosol dynamical simulations were carried out in order to gain insight into the processes

behind atmospheric particle formation (papers III–V). The advantage of an

aerosol dynamical model is that we can track the whole process of atmospheric particle

formation from nucleation size at 1–2 nm up to 500 nm under controlled conditions.

The simulations were performed using the University of Helsinki Multicomponent

Aerosol model (UHMA), which has been developed at the University of Helsinki by

Korhonen et al. (2004) and designed specifically for studies of new particle formation.

UHMA is a box-model to simulate the dynamics of the aerosol population in a uniform

”box” of air, without any advection or turbulent transport fluxes. It calculates the evolution

of the aerosol size distribution under all the basic aerosol dynamical processes

for clear-sky conditions: nucleation, condensation, coagulation and dry deposition.

The two main approaches in atmospheric modelling are the Eulerian and Lagrangian

frameworks. In the Eulerian framework, a situation at stationary observer is modelled,

whereas in the Lagrangian framework, a situation for an observer moving with the

air flow is considered. In principle, as UHMA does not include advective mass transfer,

it represents a Lagrangian modelling framework. The measurements e.g. at the

Hyytiälä SMEAR II station, however, are done at a stationary point i.e. in an Eulerian

perspective. The results from a Lagrangian box-model can be compared to measurements

at a stationary station, provided that air flow at the station is from the same

direction for a sufficiently long time, and no strong horisontal transport occurs (calm

conditions). In that case the station is observing the same and rather homogenous air

mass. This condition is often fulfilled for the class I events (Dal Maso et al., 2005; an

example shown in Fig. 2). In the steady conditions of one air mass, the observed new

particle formation behaves smoothly, having a steady growth without abrupt changes

in the size distribution.


Figure11: ThestructureoftheUHMA(University ofHelsinki Multicomponent Aerosol

model). The model is a 1-dimensional box-model designed for particle formation studies.

ThebasicstructureoftheUHMAmodelispresentedinFig. 11. Asaninput, themodel

takestheinitialparticlesizedistributionandcomposition. Inaddition, thecondensable


The model then calculates the dynamics of the aerosol population, subject to the four

aerosol dynamical processes, according to the General Dynamic Equation (Eq. 4). In

the basic version of UHMA, the integration of differential equations is performed with

the simple, 1 st order Euler-forward method. At chosen time intervals, e.g. 10 min., the

model gives as an output the current particle size and composition distribution as well

as the concentrations of condensable gases.

UHMA is a sectional model, meaning that the size distribution is divided into sections

uniformly distributed on a logarithmic particle diameter axis. The number of sections

can be chosen freely according to the accuracy needed in the study; typically 40–60

sections between 1–500 nm are used. Particles are assumed to be totally internally

mixed within a size section, i.e. all particles in a size bin have the same composition.

The particles are composed of the following ”substances”: sulphuric acid, 1–3 organic

compounds, a possible insoluble core, ammonia and water.

Condensation is calculated using the transition-regime theory of Fuchs and Sutugin


(1971) (also in Seinfeld and Pandis, 2006), with the modification by Lehtinen and

Kulmala (2003) for the molecular regime condensation flux. For condensation calculations,

UHMA applies a hybrid-sectional method, in which the particle is divided into

a ”core particle” part including sulphuric acid, organic compound(s) and possible nonsoluble

core-particle, and a ”non-core part” including water and ammonia. For the

”core” part the condensation (or evaporation) flux and the resulting diameter increase

(or decrease) is calculated explicitly from the condensation equations, whereas water

and ammonia uptake are calculated through an equiliblium parameterisation, which

depends on relative humidity, ammonia concentration, particle size and composition

(Napari et al., 2006). One of the condensing organic compounds can be set to be

a ”nano-Köhler”-compound, meaning that its condensation follows the nano-Köhlermechanism

proposed by Kulmala et al. (2004a) (see Sect 2.1.1). If more than one

condensable organic compounds are used, the rest are assumed to be water-insoluble

and follow the normal Kelvin effect.

Coagulation is calculated according to the conventional equations by Fuchs (1964,

also in Seinfeld and Pandis, 2006). Dry deposition follows a parameterisation for

Hyytiälä conditions by Rannik et al. (2003). For nucleation, several different parameterisations

can be used: binary, ternary, activation and kinetic nucleation mechanism.

To update the model for this study, I added the subroutines for activation and kinetic

nucleation mechanisms (Sect. 4.2, Eqs. 46 and 47) as well as a subroutine for sulphuric

acid production rate. The sulphuric acid production rate is calculated as a chemical

reaction rate of SO 2 and OH, while for OH a sinusoidal profile dependent on the zenith

angle of the sun was assumed, correspoding to cloudless conditions.

In this study, the following model set-up was applied (for the base case): Particles

consist of sulphuric acid, water, ammonia and one nano-Köhler-organic compound.

The properties of the organic compound were chosen to correspond to a possible VOC

oxidation product with a saturation vapour concentration c sat ≤ 10 6 cm −3 (c sat corresponds

to saturation vapour pressure via ideal gas law) (Kulmala et al., 1998b, 2001b).

The sulphuric acid saturation concentration is assumed to be zero, i.e. it condenses

with the maximum flux. For the activation and kinetic coefficients the mean values

determined for the QUEST II campaign (Hyytiälä) were used, namely A = 1 × 10 −6

s −1 and K = 5×10 −13 cm −3 s −1 (paper I).

The UHMA model (and its modified versions) has been applied widely in studies of

atmospheric aerosols and particle formation: in combination with a chemistry model

for new particle formation studies (Grini et al., 2005); in a pseudo-Lagrangian way for

simulating aerosol transformation during continental transport, with a simple treatment

for boundary layer entrainment/detrainment (Komppula et al., 2006; Tunved

et al., 2006b); in studies of iodine oxide induced nucleation in coastal, marine environment

(Vuollekoski et al., 2009; Ehn et al., 2010); for investigating sea salt aerosol

and its effect on marine aerosol and cloud droplet number concentrations (Mårtensson

et al., 2010); and for studying cloud processing and CCN activation (Korhonen et al.,

2005). Leppä et al. (2009) have developed a model version including atmospheric

ions and charged particles. In addition, UHMA has been incorporated as part of a

1-dimensional columnar model MALTE (Boy et al., 2006; Lauros et al., 2011).


4 Connection between sulphuric acid and new particle


As described in previous sections, sulphuric acid is the key compound in atmospheric

nucleation. In addition to nucleation, sulphuric acid participates in the condensational

growth of particles. Due to its extremely low saturation vapour pressure, sulphuric

acid starts to condense already on the smallest, freshly nucleated particles.

The main part of this thesis studies the correlation of new particle formation with

sulphuric acid concentration. This correlation was studied both by analyzing field

measurement data (papers I–III) and by conducting aerosol dynamical simulations

(paper IV). This chapter describes the results published in papers I–IV.

4.1 General correlation of sulphuric acid and new particle formation

in the field data

An important background to this study are the measurements by Weber et al. (1995,

1996, 1997), which reported concurrent measurements of sulphuric acid and freshly

nucleated particle concentrations in the 3–4 nm size range at a marine site (Mauna

Loa) and a continental site (Idaho Hill) in USA. They observed the formation rate

of 3 nm particles to be correlated with the ambient sulphuric acid concentration to

the power between 1–2, a much smaller power than predicted by classical nucleation

theory. Their measurements indicated, that nucleation would be collision controlled

(McMurry and Friedlander, 1979), but with a rate about three orders of magnitude

smaller than the kinetic collision frequency of the hydrated H 2 SO 4 molecules. Weber

et al. (1997) speculated that this would be caused by ammonia needed to stabilize the

H 2 SO 4 -H 2 O clusters.

After the studies by Weber et al., few research efforts were devoted to investigating

the correlation of particle formation with H 2 SO 4 , probably due to the lack, at that

time, of mass spectrometric instruments needed for the measurement of sulphuric acid.

The correlation between sulphuric acid and particle formation rate was ”rediscovered”

after the measurements of the QUEST II campaign in Hyytiälä in 2003, where the

sulphuric acid concentration was measured continuously with a good time resolution

using a chemical ionisation mass spectrometer (CIMS) (Fiedler et al., 2005; Kulmala

et al., 2006). The QUEST II campaign was the first field campaign in Hyytiälä with

continuous sulphuric acid concentration measurements.

The correlation studies presented in this thesis started from the observation, that on

some days the number concentration of freshly nucleated particles (3–6 nm in diameter)

follows nicely the sulphuric acid concentration after some time delay (Fig. 12).

The striking similarity between these quantities strongly suggests that sulphuric acid

participates in nucleation. The time delay ∆t arises from the time needed for growth

from the nucleation size ( 1–2 nm; in papers I–II nucleation at 1 nm was assumed)

to the detection limit 3 nm of the DMPS.


[H 2

SO 4


10 4 Day of year

N 3−6

Concentration (1/cm 3 )

10 3

10 2

10 1

∆ t = 1.4 h

10 0

10 −1

84 84.2 84.4 84.6 84.8 85

Figure 12: The number concentration of 3–6 nm particles and the concentration of gas

phase sulphuric acid on 25 th March (day 84), 2003, at the Hyytiälä SMEAR II station,

an example of a pure ”activation day” with linear correlation between sulphuric acid

andN 3−6 . Thenumberconcentrationoffreshlynucleatedparticlesfollowsthesulphuric

acid concentration after a time delay of approximately 1.4 hours.

The correlations between new particle formation and gas-phase sulphuric acid concentration

were studied in detail in papers I and II. The study was based on analysis

of field data, measured during three campaigns: QUEST II campaign in 2003 in

Hyytiälä, Finland (paper I), QUEST III campaign in 2004 in Heidelberg, Germany

(paper II), and BACCI/QUEST IV campaign in 2005 in Hyytiälä (paper II). These

data sets offered good material to investigate the connection between new particle formation

(abbreviated hereafter as NPF) and sulphuric acid in different environments,

Hyytiälä representing a ruralsite inboreal forest environment andHeidelberg a slightly

more polluted site in Central Europe.

We studied the correlation with sulphuric acid separately for the number concentration

of 3–6 nm particles (N 3−6 , obtained from the lowest four channels of the DMPS

measurements), for the formation rate of 3 nm particles (J 3 , calculated from DMPS

data by Eq. 36) and for the formation rate of 1 nm particles (J 1 i.e. the nucleation

rate, calculated by Eq. 40). In all data sets, both the number concentration N 3−6

and the particle formation rates (J 3 and J 1 ) were observed to correlate with sulphuric

acid concentration to the power of 1–2 (see Fig. 13). The correlations were very similar

in all three data sets, suggesting that the nucleation mechanism is similar in both

environments, Hyytiälä and Heidelberg.

To examine the correlations in more detail, we determined for each day the exponent

maximizing the correlation coefficient, separately for the three relationships: N 3−6 ∼

[H 2 SO 4 ] n , J 3 ∼ [H 2 SO 4 ] n andJ 1 ∼ [H 2 SO 4 ] n . These”best-fit”exponentsarecalledhere

the correlation exponents and labelled asn N3−6 , n J3 and n J1 . In all three data sets there


10 2

QUEST II Hyytiälä

QUEST III Heidelberg

QUEST IV Hyytiälä

10 2

J 3

(cm −3 s −1 )

10 0

10 −2

J 1

(cm −3 s −1 )

10 0

10 −2

10 −4

line with

slope 1

line with

slope 2

10 −4

QUEST II Hyytiälä

QUEST III Heidelberg

QUEST IV Hyytiälä

10 4 10 5 10 6 10 7

H 2

SO 4

(cm −3 ) (delayed by ∆t)

10 4 10 5 10 6 10 7

H 2

SO 4

(cm −3 )

Figure 13: Formation rate of 3 nm particles (J 3 ) (left) and 1 nm particles (J 1 ) (right)

versus sulphuric acid concentration during the three QUEST campaigns in Hyytiälä,

Finland, and in Heidelberg, Germany. The straight lines correspond to the linear and

squaredrelationship between J dp and[H 2 SO 4 ](withslopes1and2onlogarithmicaxis).

were pure exponent n = 1 days (linear correlation with [H 2 SO 4 ]) and pure exponent

n = 2 days (squared correlation with [H 2 SO 4 ]), as well as variants between n = 1–2

(Table 1). The correlation exponents could be different for N 3−6 , J 3 andJ 1 , typically in

the order that n N3−6 ≤ n J3 and n N3−6 ≤ n J1 . Based on simple theoretical calculations,

the change in the correlation exponent was attributed to sulphuric acid participating

in the growth of nucleated clusters (paper II). This conclusion was supported also by

the results from aerosol dynamical simulations presented in paper IV.

Thetimedelaybetweentherisein[H 2 SO 4 ]andN 3−6 canbeusedtoestimatethegrowth

rate of nucleated clusters from the nucleation size at 1 nm to 3 nm: GR 1−3 = 2nm/∆t

(Fiedler et al., 2005). The mean values for the initial particle growth rates were 1.2–3

nm/h in Hyytiälä and 1.3 nm/h in Heidelberg.

The strong correlation of particle formation rates and number concentration with the

H 2 SO 4 concentration implies that sulphuric acid is participating in nucleation and/or

growth of freshly nucleated particles. However, the correlation analysis does not give

ultimate proof that sulphuric acid is the nucleating compound. In principle, there

are three possibilities which could produce the observed correlation between NPF and

sulphuric acid:

(i) sulphuric acid participates only in nucleation (= formation of stable clusters

around 1–1.5 nm);

(ii) sulphuric acid participates in initial particle growth, but nucleation happens by

other substances (e.g. organic compounds);

(iii) sulphuric acid participates both in nucleation and initial particle growth.


Table 1: The exponents of the correlation N 3−6 ∼[H 2 SO 4 ] n (R is the mean correlation

coefficient) and the median values of the nucleation coefficients A and K for the

QUEST II–QUEST IV campaigns (paper II).


Mar 18–Apr 9, 2003 Feb 28–Apr 4, 2004 Apr 5–May 16, 2005

Hyytiälä Heidelberg Hyytiälä

n≈1 6 (38%) 6 (60%) 9 (45%)

n≈1.5 4 (25%) 3 (30%) 2 (10%)

n≈2 5 (31%) 1 (10%) 6 (30%)

n≈2.5–3 1 (6%) – 3 (15%)

mean R 0.85 0.75 0.82

median A (1/s) 1.0e-06 1.1e-05 2.4e-07

median K (cm 3 /s) 4.5e-13 3.9e-12 3.2e-14

Of these, the possibility (iii) is the most probable one. Option (i) can be ruled out,

because due to its small saturation vapour pressure, sulphuric acid always condenses

on particles. Therefore, if sulphuric acid participates in nucleation, it will also participate

in the growth of freshly nucleated particles. Overall, the participation of organic

compounds in nucleation or early growth is probable, e.g. the presence of organic acids

has been observed to enhance nucleation in laboratory (Zhang et al., 2004).

4.2 Activation and kinetic nucleation mechanisms

None of the previously presented nucleation theories — classical binary H 2 SO 4 -H 2 O or

ternary H 2 SO 4 -NH 3 -H 2 O nucleation — is able to explain the small correlation exponents

between new particle formation rate and [H 2 SO 4 ]. Classical binary and ternary

nucleation theories would predict correlation exponents of ∼4–10, and based on the

nucleation theorem (Eq. 14) this would mean that there are 4–10 H 2 SO 4 molecules in

the critical cluster. The correlation exponents 1–2 observed in the field measurement

data (papers I and II) are far below this.

To explain the observed linear relationship between new particle formation and sulphuric

acid concentration, Kulmala et al. (2006) proposed a new nucleation mechanism,

”activation nucleation”, in which the nucleation rate is directly proportional to

the sulphuric acid concentration:

J act = A [H 2 SO 4 ]. (46)

Here A is an empirical activation coefficient (units 1/s), which will be determined


according to measurement data.

In activation nucleation, nucleation is thought to happen via activation of small (∼1

nm) clusters, which after activation reach the critical radius and start to grow larger by

condensation of sulphuric acid and other vapours available. Two possibilities for this

activation process have been suggested: (i) small clusters which contain one sulphuric

acid molecule are activated via heterogenous nucleation of some other substance or by

surfacechemical reactions; or(ii)smallclustersofunspecifiedcomposition(e.g. organic

clusters) are activated when a sulphuric acid molecule hits them. These both processes

would generate a linear relationship between the nucleation rate and the sulphuric acid

concentration. Even though at present the theory of activation nucleation is somewhat

ambiguous, it provides a simple parameterisation that can be tested and utilized in

modelling nucleation. The physical and chemical details of the nucleation process

are lumped together in the activation coefficient, which so far is merely an empirical


To explain the squared relationship between new particle formation and sulphuric acid,

the ”kinetic nucleation” scheme was proposed (paper I), in which the nucleation rate

is proportional to the square of H 2 SO 4 concentration:

J kin = K [H 2 SO 4 ] 2 , (47)

where K is an empirical kinetic coefficient (units cm 3 /s). This nucleation mechanism

has the functional form of collision-limited kinetic nucleation of sulphuric acid, where

stable clusters are formed by collision of two H 2 SO 4 molecules (McMurry and Friedlander,

1979). However, here the prefactor K is kept as a free empirical parameter, which

will be determined based on measurement data. Similarly as in activation nucleation,

the kinetic coefficient K contains the details of the nucleation process: specifically the

probability that a collision of two sulphuric acid containing molecules/clusters results

in the formation of a stable cluster. The upper limit for K is set by the collision frequency,

which is obtained from the kinetic gas theory (Eq. 15). For H 2 SO 4 molecules

at T = 293K, the kinetic collision frequency is about 3·10 −10 cm 3 s −1 .

In papers I and II these nucleation mechanisms were examined and the values of

the nucleation coefficients A and K were determined for the first time. We fitted J 3

and J 1 with the sulphuric acid concentration and determined values for the nucleation

coefficients which produced the best fit (see Figs. 5 and 6 in paper I). The fitting

was performed separately for each day, thus obtaining a distribution of daily A and

K coefficients. Due to scatter in the data, on many days it was hard to decide which

nucleation mechanism (activation or kinetic) fitted better. Therefore, for every day the

values of both coefficients were determined.

Themedian values oftheactivationandkinetic coefficients foreach campaignarelisted

in Table 1. The values of the activation and kinetic coefficients showed large variation

from day to day: inside one data set (each the length of a couple of months) the

nucleationcoefficients variedover 1–2ordersofmagnitude. ThevaluesofAandK were


about an order of magnitude higher in Heidelberg as compared to Hyytiälä, probably

related to Heidelberg being more affected by anthropogenic pollution such as ammonia

whichcouldenhancenucleation. Thesulphuric acidconcentrationsweremeasured with

the same method (CIMS, see Sect. 3.1) during all the campaigns, but the individual

instruments were different. This causes some uncertainty to the comparison of the data

sets, as there may be some offset between the different instruments.

The variations of the A and K coefficients during different measurement campaigns

are compared in Figure 14, with also the values reported by Nieminen et al. (2009)

for EUCAARI 2007 campaign in Hyytiälä and by Paasonen et al. (2009) for Hohenpeissenberg

station in Germany included. The nucleation coefficients are similar in

magnitude for Hyytiälä and Hohenpeissenberg, although some variations between the

three Hyytiälä data sets is observed. The values of A and K are about an order of

magnitude higher in Heidelberg as compared to Hyytiälä. In Hohenpeissenberg data

set the variation of the coefficients is much larger than in Hyytiälä and Heidelberg:

over 3 orders of magnitude for A and up to 5 orders of magnitude for K. This is probably

related to the Hohenpeissenberg data set being 1.5 years long, thus covering all

seasons, while measurements in Hyytiälä and Heidelberg were from spring-summertime


In comparison with the earlier results by Weber et al. (1995, 1996, 1997), our study

shows strikingly similar results. Weber et al. (1996) reported a prefactor φ to the

kinetic collision frequency of sulphuric acid-water clusters of φ = 0.001 for Mauna

Loa and φ = 0.003 for Idaho Hill, with which the nucleation rate would be: J nuc =

φK kin [H 2 SO 4 ] 2 . With K kin = 3 · 10 −10 cm 3 s −1 , Weber’s prefactor (corresponding to

our nucleation coefficent K) would be 3–9 ·10 −13 cm 3 s −1 , which is in the same range

with our values for the kinetic coefficient in Hyytiälä.

The physical and chemical details of the nucleation prosess are hidden behind the empirical

activation and kinetic coefficients. The large variation in A and K implies that,

besides sulphuric acid, there are other factors that affect the atmospheric nucleation

rate crucially.

The existence of pure ”activation days” with linear correlation (e.g. the day in Fig.

12) and pure ”kinetic days” with squared correlation with [H 2 SO 4 ] (e.g. the day in

Fig. 3b of paper II), implies that there are different (at least two) nucleation/initial

growth mechanisms working on different days with varying ambient conditions. The

conditions affecting the nucleation mechanism can be the level of [H 2 SO 4 ], presence of

organic compounds (oxidation products of VOCs), ammonia or amine concentrations,

temperature, relative humidity etc.

After papers I and II, a few studies on the connection of NPF with sulphuric acid

have been published. Kuang et al. (2008) investigated the sulphuric acid correlations

at several locations by applying somewhat different methods for J 3 calculation (Eq.

31) and correlation analysis than in this thesis. They found that kinetic nucleation

(with exponent very close to 2) was explaining nucleation in all studied places. In their

analysis, Kuang et al. combined the data points from different days into one data set,


A (s −1 )

10 −4

10 −5

10 −6



25 % − 75 %

Min − Max

K (cm 3 s −1 )

10 −10

10 −12



25 % − 75 %

Min − Max

10 −7

10 −14

10 −8

Hyy Q2 Hyy Q4 Hyy 2007 Hei Q3 Hohenp

10 −16

Hyy Q2 Hyy Q4 Hyy 2007 Hei Q3 Hohenp

Figure 14: The values of activation (A, left) and kinetic (K, right) coefficients during

five measurement campaigns: QUEST II, QUEST IV and EUCAARI 2007 (Nieminen

et al., 2009) at the Hyytiälä SMEAR II station; QUEST III in Heidelberg, Germany;

and HAFEX at the Hohenpeissenberg station, Germany (Paasonen et al., 2009).

for which the correlations were determined, whereas in the studies of this thesis the

correlations were investigated separately for each day.

Paasonen et al. (2009, 2010) examined the effect of organic vapours in activation and

kinetic nucleation mechanisms. The nucleation coefficients A and K were observed

to correlate positively with monoterpene oxidation products, but no such correlation

existed for the nucleation rate (J 1.5 ) (Paasonen et al., 2009). A comprehesive study of

data from Hyytiälä, Hohenpeissenberg (Germany), Melpitz (Germany) and San Pietro

Capofiume (SPC, Italy) showed that overall, kinetic sulphuric acid nucleation explains

nucleation well in all other places than Hohenpeissenberg, where the effect of organics

is dominant. However, even at the sulphuric acid-dominated sites (Hyytiälä, Melpitz,

SPC),theinclusionofco-nucleationoforganicvapourandsulphuric acid(togetherwith

kinetic nucleation of H 2 SO 4 ) improved the nucleation rate prediction (Paasonen et al.,

2010). The importance of organic vapours very probably explains the high variation

for A and K in Hohenpeissenberg shown in Fig. 14.

There are indications that ammonia or amines (Petäjä et al., 2011; Zhao et al., 2011;


for atmospheric nucleation. Brus et al. (2011) have reported laboratory measurements

of H 2 SO 4 -H 2 O nucleation, in which the correlation exponent was observed to decrease

with increasing temperature (from n = 2.2 at 5 ◦ C to n = 1.2 at 25 ◦ C).

In papers I–II the exponent of the correlation J 1 ∼ [H 2 SO 4 ] n (the slope of the log(J 1 )

vs log([H 2 SO 4 ]) plot) was interpreted as the number of H 2 SO 4 molecules in the critical

cluster, based on the approximate version of the nucleation theorem (Eq. 14). According

to current knowledge, this conclusion does not hold: if the Gibbs free energy curve

haslocal minima, thebasic nucleation theorem isnot valid, andat least theapproximation

n ∗ +1 ≈ n ∗ can not be done, when n ∗ is small. However, the correlation exponent

can be interpreted as giving information on the rate limiting step in atmospheric NPF

(cluster formation or their growth): this step seems to be proportional to the H 2 SO 4


concentration to the power between 1 and 2.

When papers I and II were published, there existed a remarkable gap between atmospheric

and laboratory measurements of nucleation (see Sect. 2.2.3): the slope of

log(J nuc vs log([H 2 SO 4 ]) curves were significantly higher in the laboratory (order of

4–10) than in the atmosphere. This discrepancy has been overcome by recent advances

in measurement technologies of 1–3 nm particles, and the slopes of 1–2, as

observed in atmospheric measurements, have been reproduced also in laboratory conditions

(Sipilä et al., 2010).

Korhonen et al. (2010) made a computational study on the sulphuric acid correlations,

investigating the accuracy of the analysis methods to determine the exponent of nucleation

andthe values of the nucleation coefficients (A and K). The result was that there

are several uncertainties in the analysis procedure, especially in the determination of

nucleation rates from the particle formation rates at 3 nm. Therefore, the correlation

exponents and values of the nucleation coefficients determined with several different

steps in the data analysis procedure should be interpreted with caution: at least noting

that they contain significant error bars.

Despite their deficiencies (the large variation of A and K even within the same data

set), thedeveloped parameterisations foractivationandkinetic nucleation have already

been used quite widely in globalaerosol andclimate models (e.g. Spracklen et al., 2006,

2008; Makkonen et al., 2009). The parameterisations capture quite well nucleation

happening in the atmospheric boundary layer.

Summarizing, the atmospheric particle formation rate is observed to depend on the

sulphuric acid concentration to the power 1–2 (papers I and II). According to current

knowledge, the kinetic nucleation (with exponent 2) seems to be most widely valid,

together with co-nucleation of sulphuric acid and organics. In some places, nucleation

may be dominated by organic compounds (Paasonen et al., 2010). Pure activation

nucleation (with linear correlation with H 2 SO 4 ) seems to be a special case, which

certainly acts in some conditions, but overall it happens rather rarely. In the future,

aerosol mass spectrometric measurements will probably reveal the constituents of the

nucleated clusters. The final goal is to develop a parameterisation and a theoretically

consistent framework that captures all important factors affecting the atmospheric

nucleation rate.


4.3 The effect of relative humidity on the nucleation rate

Atmospheric newparticleformationhasbeenknowntohappenpreferablyinconditions

with low relative humidity (RH) (e.g. Birmili et al., 2000; Boy and Kulmala, 2002;

Hyvönen et al., 2005). This has been observed consistently in several locations, but

no solid explanation for the observation has been given so far. It has been speculated,

that the main reason would be the increased coagulation and condensation sink, due

to background particles’ diameter increase when they take up water from humid air.

This leads, on one hand, to increasing coagulational scavenging of small, nucleated

particles; on the other hand, increased condensation sink decreases the concentrations

of condensable vapours, thereby hindering nucleation and initial particle growth. Both

these effects act in the same direction, and prevent particle formation.

In contrast to atmospheric observations, in laboratory measurements of nucleation in

H 2 SO 4 -H 2 O or H 2 SO 4 -NH 3 -H 2 O-system, the nucleation rates are consistently observed

to increase with relative humidity (Berndt et al., 2005; Brus et al., 2011; Benson et

al., 2008, 2009). Also from a theoretical point of view this would be expected. As

atmospheric particle formation is thought to happen via co-nucleation of sulphuric

acid and water or via a ternary mechanism involving also ammonia, increasing water

vapour concentration should enhance nucleation as the formation energy of a critical

cluster islowered whentheconcentration(orsaturationratio)ofanucleating substance

increases (Vehkamäki et al., 2002, Merikanto et al., 2007).

In paper III, the reasons behind the RH-inhibition of new particle formation were

examined in detail with the aid of field measurement data, theoretical calculations

and aerosol dynamical simulations. A new hypothesis for RH-inhibiting effect was

presented: decreased solar radiation in humid conditions might limit sulphuric acid

production and thereby lead to smaller nucleation rates. The purpose was to find out,

which of the following effects is the dominating one in preventing NPF at high RH:

i) effect of RH on H 2 SO 4 concentration via reduced OH concentration (reduced

solar radiation) at high RH;

ii) effect of RH on H 2 SO 4 concentration via increased condensation sink at high RH;


iii) effect of RH to J 3 through increased coagulation sink at high RH.

New hypothesis for the RH-inhibiting effect on NPF

Thenewhypothesiswasbasedonthefieldmeasurement dataofparticleformationrates

(J 1.5 ) and sulphuric acid concentrations presented in Fig. 15. When the points in J 1.5

vs [H 2 SO 4 ] plot were colour scaled according to relative humidity, a clear dependence

on RH emerged: nucleation rates were highest when RH was lowest and vice versa.

Whenplottedwithabsolutehumidity (water vapourconcentration), nosuch separation

of points happened as with RH.


H 2

O (%o) as colour code (QUEST II)







J 1.5

(cm −3 s −1 )





J 1.5

(cm −3 s −1 )







points between

6 a.m. − 6 p.m.

10 2

10 0

10 −2

10 −4

10 4 10 6

[H 2

SO 4

] (cm −3 )

10 8

10 4


RH (%) as colour code (QUEST II)

points between

6 a.m. − 6 p.m.

10 2

10 0

10 −2

10 −4

10 4 10 6

[H 2

SO 4

] (cm −3 )

10 8

10 4


Figure 15: Nucleation rate versus sulphuric acid (in log axes), colour scaled according

toabsolutehumidity(H 2 Oconcentration, left)andrelativehumidity(RH,right)during

QUEST II campaign at the Hyytiälä SMEAR II station. The lines correspond to the

linear and squared correlation between J 1.5 and [H 2 SO 4 ].

The observation shown in Fig. 15 led us to present a hypothesis that the RH-effect

on the nucleation rate is mediated through sulphuric acid concentration: high RH

may prevent H 2 SO 4 formation via decreased photochemistry due to decreased sunlight

reaching the ground on hazy or partially cloudy days with high RH. Decreased solar

radiation leads to decreased formation of OH radicals, which further affects the formation

of H 2 SO 4 through reaction of SO 2 and OH (see Sect. 2.3). As the nucleation rate

is controlled by sulphuric acid (as reported in papers I and II), smaller sulphuric acid

concentrations are directly transferred to smaller nucleation rates.

Data analysis revealed that at RH > 60 % the sulphuric acid concentrations decreased

with increasing RH. Especially the highest H 2 SO 4 concentrations were totally missing

at RH > 60 % (see Fig. 1 in paper III). The SO 2 concentration was observed to

be rather independent of RH. Instead, OH and UV showed very similar correlation

with RH as H 2 SO 4 : both the OH concentration and UV radiation intensity started

to decrease above a RH of 60 %. Taken together, this data suggests that RH limits

nucleation through limiting UV-B and OH, and thereby sulphuric acid production.

The decreasing effect of RH on UV radiation can be explained as follows. At humid

conditions, the probability of cloud and fog formation increases, and in the presence

of clouds the UV radiation reaching the lower atmosphere (boundary layer) decreases.

The relationship between cloudiness and RHwas investigated using a 30-year long data

set oflow level clouds collected by theFinnish Meteorological Institute at theJokioinen

weather station. As expected, the amount of low level clouds was negatively correlated

with RH. Interestingly, there seemed to be a threshold value of RH (about 40–60 %)

above which cloudiness started to increase steeply with RH, especially around midday

hours (see Fig. 5 in paper III). A plot of global radiation versus RH in the Jokioinen

data set showed a similar decreasing trend with RH as in the Hyytiälä QUEST II data

set, and a mirror-like behaviour as compared to cloudiness. It is worth noting, that a


H 2

O (%o) as colour code (EUCAARI 2007)

points between

6 a.m. − 6 p.m.



10 4

RH (%) as colour code (EUCAARI 2007)

points between

6 a.m. − 6 p.m.



10 2


10 2


J 1.5

(cm −3 s −1 )

10 4 [H 2

SO 4

] (cm −3 )

10 0

10 −2





J 1.5

(cm −3 s −1 )

10 0

10 −2







10 −4

10 4 10 6 10 8


10 −4

10 4 10 6 10 8

[H 2

SO 4

] (cm −3 )


Figure 16: Nucleation rate versus sulphuric acid, colour scaled according to absolute

humidity(H 2 Oconcentration, left)andrelativehumidity(RH,right)duringEUCAARI

2007 campaign at the Hyytiälä SMEAR II station. The lines correspond to the linear

and squared correlation between J 1.5 and [H 2 SO 4 ].

similar threshold at RH = 40–60 % was observed for all variables: H 2 SO 4 , OH, UV,

global radiation and cloudinesss. The cloudiness data supports the hypothesis that the

RH limits the H 2 SO 4 concentration through limiting OH production, and this effect

is because of decreased UV radiation reaching the ground due to cloudiness in humid


The study presented in paper III was based on data from the QUEST II campaign

from Hyytiälä. The similar correlation of J 1.5 and H 2 SO 4 with RH (but not with H 2 O

concentration) is seen also for QUEST IV and EUCAARI 2007 data sets (see Fig.

16) from Hyytiälä in spring 2005 and 2007, respectively. This gives confidence that

the results are more generally valid, at least in conditions at the Hyytiälä SMEAR II


Another reason for the observed anti-correlation of OH and UV with RH could be,

that these variables have opposite diurnal profiles: UV radiation (and thus OH) peaks

at noon, whereas RH has minimum at noon/afternoon, when temperature and water

saturation concentration are the highest. This results in an apparent anticorrelation

between UV (and OH) with RH, without a necessary causal correlation between these

variables. The effect of opposite diurnal variations makes it difficult to distinguish how

much of the observed correlation is explained by cloudiness and how much is only due

to diurnal variations. In paper III it is suggested that opposite diurnal cycles would

be the main reason for the anticorrelation of OH and UV (and H 2 SO 4 ) with RH, and

the effect of cloudiness would be a minor effect.


Effect of increased condensation and coagulation sink

At humid conditions aerosols experience hygroscopic growth: they take up water,

thereby increasing their diameter and surface area. The DMPS measures the particle

size distribution as a dry distribution at about RH = 30 % prevailing inside the

instrument. The hygroscopicity of particles then has to be taken into account using

a parameterisation for particle wet diameter, in this case with a parameterisation by

Laakso et al. (2004) for Hyytiälä conditions. As an example, from RH of 30 % to 90 %

hygroscopic growth increases the condensation and coagulation sinks by about a factor

of 3 (paper III).

The increased condensation sink (CS) increases the loss rate of sulphuric acid, leading

to smaller H 2 SO 4 concentrations. This is further transferred to a smaller nucleation

rate and particle formation rate at 3 nm, J 3 . The increase in coagulation sink (CoagS),

in turn, increases the coagulational scavenging of small clusters. The probability of a

cluster surviving from the nucleation size (1–1.5 nm) to 3 nm size becomes smaller,

leading to a smaller particle formation rate at 3 nm (J 3 , as described by Eq. 37 or 43).

The effect of increased coagulation sink was examined by theoretical calculations performed

using the new form of the Kerminen-Kulmala equation, Eq. 43 (Lehtinen et

al., 2007). We calculated how much J 3 decreases from the nucleation rate J 1.5 at different

relative humidities, when wet-CoagS is calculated with the parameterisation by

Laakso et al. (2004). The result was that increased CoagS could suppress new particle

formation, provided that the particle growth rate is low and the coagulation sink

(background aerosol concentration) is high (”extreme case” in paper III). In such

conditions, J 3 decreased by 1–3 orders of magnitude from J 1.5 when RH increased from

10 to 90 %, indicating that at high RH CoagS could mask the new particle formation

event from observation at 3 nm. However, at average conditions of typical growth rate

and CoagS, the decrease from J 1.5 to J 3 was smaller, and not enough to suppress the

nucleation event.

TherelativeimportanceoftheeffectsofdecreasedOHproduction, increasedCoagSand

increased CS were further investigated by performing aerosol dynamical simulations.


%, based on the observed correlation of OH with RH (Fig. 4 in paper III) and

analysis from a chemical model (Boy et al., 2005). The modelling results showed that,

of these variables, the particle formation rate J 3 was most sensitive to the reduced

OH levels, as decreased OH directly leads to smaller sulphuric acid concentrations and

nucleation rates. The effects of increased CS (through decreased H 2 SO 4 concentration)

and increased CoagS (through the coagulational scavenging of small clusters) on J 3

were similar in magnitude, but these effects were significantly smaller than that of

reduced OH concentration.

In conclusion, paper III reports the finding that RH seems to limit the nucleation rate

by limiting OH and H 2 SO 4 production at high relative humidities. In comparison to

OH-effect, the earlier proposed mechanisms for the RH-inhibiting effect, increased condensation

and coagulation sinks, are shown to have smaller contribution in suppressing


new particle formation. The effect of RH on OH could be related to cloudiness at

high RH: cloud and haze droplets increase the scattering of sunlight, leading to less

radiation reaching the ground and smaller OH production rates. However, it is possible

that the observed anticorrelation of H 2 SO 4 and RH is mainly due to opposite

diurnal profiles of UV radiation and RH. Paper III concludes that ”even though at

first glance RH appears to limit NPF, this appearance is due to its anticorrelation with

solar radiation.”

The study of paper III demonstrates how difficult it is to distinguish real reasons and

causal relationships behind an observed correlation in atmospheric data, when almost

all variables have a diurnal variation related to sunlight. Actually, paper III provokes

even more questions than it answers. A continuing study would be needed, with more

sophisticated data analysis methods to separate the diurnal variation from the other

variations in the data. Also, it would be interesting to repeat the investigation at other

sites, to find out how generally the proposed mechanism of supression of new particle

formation by reduced OH production at high RH is valid.

4.4 Modelling the connection between sulphuric acid and particle


In papers I and II the strong correlation between new particle formation and gasphasesulphuric

acidconcentration wasstudiedbasedondatameasured atfieldstations

in Hyytiälä and Heidelberg. The activation and kinetic nucleation mechanisms were

proposed as explanations for the linear or squared relationship between particle formation

at 3 nm and sulphuric acid concentration. In order to study in more detail this

correlation, aerosol dynamical simulations were performed (paper IV). The purposes

of the simulations were:

(i) to find out whether other factors than the nucleation mechanism, such as condensation

and coagulation, affect the correlations of J 3 and N 3−6 with [H 2 SO 4 ];

(ii) to examine how the correlation with [H 2 SO 4 ] changes as particles grow from

nucleated size (1–2 nm) to 3 nm;

(iii) to find the conditions which would yield the observed linear correlation (on some

days) between particle number concentration N 3−6 and sulphuric acid.

Thesimulations were performedwiththeUHMAmodel using threedifferent nucleation

mechanisms: activation, kinetic and ternary nucleation. The condensing vapours were

sulphuric acid and an organic vapour with nano-Köhler mechanism for condensation.

Sulphuric acid had a sinusoidal profile with maximum of 5–7 ·10 6 cm −3 at noon, corresponding

to typical sulphuric acid concentrations in Hyytiälä, whereas the organic

vapour had a constant concentration of 10 7 cm −3 . The relative humidity was 50 % and

ammonia (NH 3 ) concentration 5 ppt. To investigate how the correlation of J 3 andN 3−6


with [H 2 SO 4 ] depends on the particle condensational growth rate and the size at which

nucleation is happening (objective ii), we varied the saturation vapour concentration of

the organic vapour (c sat,org , which affects the profile of the growth rate as a function of

particle diameter), and the size of the nucleated cluster (d nuc ), respectively. In total, a

high number of sensitivity runs were performed; paper IV reports in more detail the

results of six case studies representing the main features of the simulations (Fig. 17).

The correlations with sulphuric acid concentrations were studied separately for the

nucleation rate, the particle formation rate at 3 nm and the number concentration of

3–6 nm particles. The correlation exponents were determined for J nuc ∼ [H 2 SO 4 ] nnuc ,

J 3 ∼ [H 2 SO 4 ] n J 3 and N 3−6 ∼ [H 2 SO 4 ] n N 3−6 by finding the slope of the log-log curve.

Figure17: Simulated new particleformationevents applying different nucleation mechanisms:

activation (top), kinetic (middle) and ternary nucleation (bottom). The left

column (a, c, e): c sat,org = 10 6 cm −3 and nano-Köhler mechanism for the condensable

organic vapour; the right column (b, d, f): c sat,org = 0 cm −3 . The white horizontal line

shows the 3 nm border.


Table 2: The correlation exponents n J3 and n N3−6 for J 3 ∼ [H 2 SO 4 ] n J 3 and N 3−6 ∼

[H 2 SO 4 ] n N 3−6, for events simulated by the UHMA model (Fig. 17), using different nucleation

mechanisms and condensing organic vapour properties. The nucleated cluster

size was in all cases d nuc = 1 nm.

Activation nucleation Kinetic nucleation Ternary nucleation

n nuc = 1 n nuc = 2 n nuc ≈ 5.6

c sat,org = 10 6 cm −3 n J3 = 3.2 n J3 = 3.4 n J3 = 5.6

n N3−6 = 2.3 n N3−6 = 2.3 n N3−6 = 4.1

c sat,org = 0 cm −3 n J3 = 1.3 n J3 = 2.1 n J3 = 5.0

n N3−6 = 1.2 n N3−6 = 1.7 n N3−6 = 4.0


for these three quantites (Table 2). Especially with activation nucleation, which has

nucleation exponent n nuc = 1, the correlation exponent could increase by 1–2 units for

J 3 and N 3−6 . Generally, the exponent for N 3−6 correlation was smaller than for J 3 .

In case of ternary nucleation, the correlation exponent for N 3−6 was even smaller than

the nucleation exponent.

These changes in sulphuric acid correlation exponents may seem peculiar. However,

they are explained by aerosol dynamical processes happening during the growth from

nucleation size at 1–2 nm to 3–6 nm. The simulations revealed that most important

for the change of the correlation exponent is the particle growth rate between 1–3

nm, which is both size and time dependent. The size dependence of the growth rate

with different values of organic vapour saturation concentration (c sat,org ) is presented

in Fig. 18. At high saturation concentration (c sat,org = 10 5 –10 6 cm −3 ), the growth of

the smallest particles (1–1.5 nm) is solely due to sulphuric acid, and the organic vapour

starts to condense gradually between 1.5–4 nm, following the nano-Köhler mechanism

(see Fig. 5 in paper IV). When the organic saturation concentration decreases, the

growth rateof small particles increases, as organicvapour starts to condense onsmaller

and smaller particles. At c sat,org = 0, the organic vapour condenses with the maximum

flux (not limited by the Kelvin effect).

With activation nucleation, preserving the exponent of nucleation for J 3 and N 3−6

required either i) fast growth of the nucleated particles, or ii) nucleation happening at

2 nm size. For the first requirement, the nano-Köhler mechanism (with non-negligible

saturation concentration) for organic vapour condensation was not alone sufficient to

increase the growth rate of the smallest particles, but a negligible saturation (c sat,org ≤

10 2 cm −3 ) concentration was required to get high enough growth rate (see Fig. 18). In

both these cases, particles grow fast from nucleation size to 3–6 nm, so that also the

sulphuric acid correlation ”has no time to change much”.

In addition to condensational growth, the coagulation loss of freshly nucleated particles

affects the correlation of J 3 and N 3−6 with sulphuric acid. This question was not

addressed in the study of paper IV. As coagulation becomes important at high nucle-


ation rates, it was speculated that the different behaviour of the correlation exponents

in case of ternary nucleation (exponent decreases from J nuc to J 3 and N 3−6 ) could

be caused by the effect of coagulation (the ternary mechanism causes more intense

nucleation, as seen in Fig. 17).

Growth rate (nm/h)









GR−total at t = 12 h

c sat

= 10 6 cm −3 (Case 1)

c = 10 5 cm −3


c = 10 4 cm −3


c sat

= 10 3 cm −3

c = 10 2 cm −3


c = 0 cm −3 (Case 2)





1 2 3 4 5 6 7 8

Particle diameter (nm)

Figure 18: The particle growth rate due to condensation of sulphuric acid and organic

vapour as a function of particle size in the simulations made with the UHMA

model. The different curves are for the different values of the organic vapour saturationconcentration(c

sat,org ). Theorganicvapourcondensationfollowedthenano-Köhler


Asaconclusion, thecorrelationwithsulphuric acidcanchangesignificantly during particle

growth from the nucleation size 1–1.5 nm to 3–6 nm, and therefore the correlation

exponents observed for J 3 or N 3−6 should not be interpreted directly as the exponent of

nucleation. However, modelling results showed that the correlation exponent can not

change very much, only by 1–2 units, and therefore ternary nucleation, having a correlation

exponent 4–6, can be most probably ruled out, at least for nucleation happening

in the boundary layer in boreal forest conditions. According to the results presented

in this thesis, one can safely say that the nucleation mechanism seems to involve sulphuric

acid and has a dependence on the sulphuric acid concentration to the power of

1–3. This result is supported also by laboratory studies with novel instrumentation

(Sipilä et al., 2010; Brus et al., 2011).


5 CCN activity of boreal forest aerosols

Aerosols are needed in cloud formation as seeds for water condensation, i.e. to act

as cloud condensation nuclei (CCN). An aerosol particle can act as a CCN, if its size

is bigger than the critical (threshold) diameter for irreversible water condensation to

happen. In the fast condensation of water, called activation, the CCN are converted to

cloud droplets. The critical size depends on the physical size of the particle (diameter)

and its chemical composition (solubility to water). Typically aerosol particles bigger

than 50–100 nm are potential CCN. It is probable, that a considerable fraction of

particles acting as CCN are produced by atmospheric nucleation events (Merikanto

et al., 2009; Kerminen et al., 2012).

Aerosol climate effects through acting as CCN are called the aerosol indirect effects

(Penner et al., 2004). In cloud formation, increasing CCN concentrations inside a cloud

cause a larger number of cloud droplets but smaller in size to be formed, assuming that

the same water amount is condensing on a bigger number of CCN. This has twofold

effects on cloud characteristics: i) smaller cloud droplets scatter solar radiation more

efficiently, resulting in a higher cloud albedo i.e. whiter clouds (first indirect effect)

and ii) smaller cloud droplets do not precipitate as easily, making the cloud lifetime

longer (second indirect effect). Thus, increasing aerosol concentrations imply whiter

and longer-lasting clouds, exerting a cooling effect on the climate (see Fig. 1). In

addition to these cooling effects, clouds scatter back infrared radiation coming from

the Earth’s surface. Also, black carbon and dust aerosols can form so called ”brown

clouds”, which are actually thick aerosol layers and not clouds. These brown clouds,

which are encountered in heavily polluted areas in Asia (India and China), absorb

solar radiation and via complex effects on atmospheric temperature, may have both

warming and cooling effects on the climate (semi-direct effects) (Ramanathan et al.,

2007; Koch and Del Genio, 2010).

Aerosol-cloud effects are one of the poorest understood parts of the climate system.

According to current knowledge, aerosols have during the past 50–100 years caused

a significant net cooling effect on the climate, thus masking part of the warming by

greenhouse gases (Andreae et al., 2005; Makkonen et al., 2012).

The possibility to cool the climate artificially, e.g. by inserting sulphate aerosols in

to the stratosphere (where they directly scatter sunlight) or sea-salt aerosols to the

marine troposphere (where they affect the formation of stratocumulus clouds), has

been suggested as a possible way to counteract climate warming. This is called geoengineering,

and despite the big risks associated with it, it has become a serious path

in climate research (e.g. Partanen et al. (2012)). Also for this reason, understanding

the aerosol-cloud interactions is crucial.

It has been proposed that Nature itself could also have a feedback mechanism through

aerosols which would decrease the climate warming (Kulmala et al., 2004c; Tunved

et al., 2008; Paasonen et al., 2013). Natural aerosols are to a big extent formed on

continental areas with forests and other vegetation. Formation of secondary organic

aerosols could increase in the warming climate, as most of the organic vapour emissions


increase with temperature. This would increase direct scattering of solar radiation and

affect cloud formation, thus having a cooling feedback effect on climate.

Boreal forest areas, covering 8 % of the Earth’s surface, are even a globally important

source of secondary organic aerosol (SOA) particles (Tunved et al., 2006a), which

have climatic effects at least on a regional scale (Lihavainen et al., 2009). In paper

VI the ability of boreal forest aerosols to act as cloud condensation nuclei was

investigated based on 1 year of continuous CCN concentration measurements at the

HyytiäläSMEARIIstation. Inaddition, concurrent hygroscopicitymeasurements were

analysed, to get more information on the CCN activity of aerosols in Hyytiälä.

5.1 Seasonal variation of CCN properties at the Hyytiälä

SMEAR II station

The measured CCN concentration at a certain supersaturation is determined by the

aerosol number concentration, the aerosol size distribution (whether there are more

large, CCN active particles or more small, non-CCN-active particles), and the chemical

composition(thehygroscopicity) oftheparticles. Thechemical compositiondetermines

the critical diameter for cloud droplet activation.

Measured CCN concentrations were observed to have a clear seasonal variation, with

highest concentrations inthesummertime (Fig. 2inpaper VI).InJune-JulytheCCN

concentration was at least double the concentration in December-January. Also the

activated fraction (calculated in this study as the total fraction of particles activated to

cloud droplets, F act = N CCN /N CN , where N CN is the total particle number concentration)

had a maximum in summertime. The seasonal variation of the critical diameter

was a bit different, but it also had the smallest values during the growing season in

spring-summer. These findings indicate that, on average, the aerosol particles in the

summertime boreal forest are more hygroscopic and more CCN-active than aerosols in

wintertime. In summertime, aerosols are expected to contain a larger fraction of oxidation

products of organic substances emitted from the forest; the high CCN activity

of the summertime aerosol particles in Hyytiälä suggests that these organics are highly


The critical diameter (d crit ) for cloud droplet activation was determined in this study

by two different methods. First, d crit was determined from CCN concentrations and

particle size distribution (DMPS) data by simply integrating the size distribution from

the largest particle size towards smaller sizes, until the concentration matches the

measured CCN concentration:



d i =d crit

N i = N CCN , (48)

where the lower size limit corresponds the critical diameter. In this method it is

assumed that particles are internally mixed (i.e. particles of certain size have the same


Figure 19: Average critical diameters as a function of water supersaturation for

Hyytiälä aerosol, estimated from CCNC and HTDMA data (error bars indicate the

standard deviations in the one year data set). Laboratory measurements for secondary

organic aerosol (α-pinene and trimethylbenzene, TMB) are shown for comparison (Duplissy

et al., 2008). Lines show the theoretical values from the kappa-Köhler equation

for pure ammonium sulphate particles and for a range of SOA kappa values.

chemical composition), all particles larger than d crit are activated (i.e. differences in

chemical composition are neglected) and the activation probability behaves as a step

function at d crit .

Second, the critical diameter was estimated from the hygroscopicity measurements

(HTDMA) using kappa-Köhler theory (see Sect. 2.4). The value of the hygroscopicity

parameter κ (kappa) was determined from HTDMA-data (for conditions S < 1), and

after that the value of the critical (dry) diameter was calculated from Eq. 29 for the

supersaturations corresponding to CCN measurements (SS = 0.1–1.0 %).

The critical diameters determined by these two methods corresponded well to each

other, the mean values being ∼50 nm, ∼80 nm and 150–200 nm at SS of 1.0 %, 0.4

% and 0.1 %, respectively. While there were differences between daily and monthly

values, on average both methods produced very similar behaviour of d crit as a function

of supersaturation (see Fig. 19 and Table 1 in paper VI). The points fall nicely

on the theoretical slope of −2/3 on the log-log plot (d crit ∼ SS −2/3 ). The determined

critical diameters arenotfar fromthelaboratoryresults forα-pinenesecondary organic

aerosol. In Hyytiälä, α-pinene is one of the main monoterpenes emitted by the forest.

The presence of sulphate in Hyytiälä aerosols makes the critical diameters somewhat

lower than for pure organic aerosol.


The mean kappa parameter, as determined from HTDMA measurements, was 0.18.

This corresponds to estimated fractions of 84 % organics (κ = 0.1) and 16 % sulphate

(κ = 0.6). These fractions are in line with other estimates of the organic fraction

of Hyytiälä aerosols (Boy et al., 2005; Jimenez et al., 2009). Of course, this division

to only two consituents (organics and sulphate) is very rough. There are certainly

also less-hygroscopic organics (with κ < 0.1) or purely non-hygroscopic aerosols such

as black carbon present in Hyytiälä aerosol. The value of 0.18 is an average over all

supersaturations (0.1–1.0 %), and there are most probably differences in the κ values

between different particle sizes. This is reflected also in Fig. 19: at high supersaturations

the points start to deviate from the same straight line with slope −2/3.

The value determined for the κ-parameter (κ = 0.18) can be utilised in modelling the

concentrations of potential CCN in boreal forest. To capture the seasonal variation

of CCN concentrations, a seasonal profile for κ would be needed. This remains to be

determined in future studies.

The comparison of the two methods for determination of the critical diameter showed,

that even the simplified method of integrating the size distribution is capable of giving

at least a rough estimate of the treshold diameter for CCN activation. This is in line

with earlier results, which report that the particle size distribution dominates over

the chemical composition in determining the CCN concentrations (Dusek et al., 2006;

Quinn et al., 2008). In principle, the simplified integration method should give an

upper limit to d crit , because all particles larger than that size are assumed to activate,

regardless their chemical composition.

5.2 Effect of new particle formation on cloud condensation


Figure 20 shows an example of a period with several new particle formation (NPF)

events, together with the measured CCN concentrations, the activated fraction and

critical diameter (determined from the CCN-concentration and particle size distribution

data by Eq. 48). The CCN concentrations increase constantly during these NPF

days, even though the total number concentration (CN) is approximately constant.

This indicates that NPF events are producing significant numbers of particles to the

CCN-active size range. The increase in CCN concentration is more pronounced for the

three highest supersaturations (SS ≥ 0.4 %), which (due to smaller critical diameters)

are more affected by the growing nucleation mode.

The effect of new particle formation on CCN concentrations was further studied by

computing the average diurnal variation of the CCN concentration, activated fraction

and critical diameter separately on new particle formation event days and non-event

days (see Figs. 7–9 in paper VI). The diurnal variation was examined for a two-day

period, because the nucleation mode typically grows to CCN active sizes (to 50–100

nm) by the end of the NPF day. Nucleation events were observed to cause an increase

in CCN concentrations, which started in the evening of the NPF event day and lasted


untiltheendofnextday(Fig. 21). AmoderateincreaseinCCNnumberwasobservable

even at the smallest supersaturations, indicating that nucleation is really capable of

producing CCN-active particles.

Diameter [m]

10 −6

10 −7

10 −8

20/04 21/04 22/04 23/04 24/04

dN/dlog(d p

) (cm −3 )

10 100 1000 10000 100000

CCN or CN [cm −3 ]

10 4

10 3

10 2

CN total


20/04 21/04 22/04 23/04 24/04

Activated fraction




20/04 21/04 22/04 23/04 24/04

SS = 0.1 %

SS = 0.2 %

SS = 0.4 %

SS = 0.6 %

SS = 1.0 %

d crit


10 2

10 1

20/04 21/04 22/04 23/04 24/04

Figure 20: Comparison of the particle size distribution data and cloud condensation

nuclei (CCN) data for a period with four consecutive new particle formation days in

Hyytiälä (20.4.-23.4.2009). (a) The time evolution of particle size distributions, (b)

CCN and total particle (CN) concentrations, (c) the activated fraction (CCN/CN),

and (d) the critical diameter estimated from CCN and particle size distribution measurements.


Elevated CCN concentrations due to new particle formation events have been reported

also in other studies (Lihavainen et al., 2003; Kuwata et al., 2008; Wiedensohler et al.,

2009; Asmi et al., 2011).

The study of paper VI was based on analysing ground-based measurements. It must

be noted, that the link between organic aerosols formed in the atmospheric boundary

layer and clouds formed in the upper part of the troposphere is not straightforward.

To really influence cloud formation, aerosols should travel from the forest up to the

height of 1–2 km. Airborne measurements are needed to find out whether the CCN

inside tropospheric clouds originate from the boundary layer, or are formed by in-situ

nucleation in the upper troposphere.


CCN concentration (cm −3 )








SS = 0.4 %

Event days

Non−event days


00:00 12:00 00:00 12:00 00:00

Time (hour of day)

Figure 21: Mean diurnal variation of CCN concentrations on two consecutive days,

compared between new particle formation event and non-event days at supersaturation

SS = 0.4 % (data points are averages over 1 year data from the Hyytiälä SMEAR II



6 Review of papers and the author’s contribution

Paper I investigates the correlation of sulphuric acid and new particle formation by

analysing field data measured during the QUEST II campaign in Hyytiälä, at the

SMEAR II station. The paper reports day-specific values for empirical activation and

kinetic nucleation coefficients for the first time. I made all the data analysis and was

responsible for writing the main parts of the article.

Paper II continues to study the correlation between sulphuric acid and new particle

formation utilizing data sets from the QUEST III and IV campaigns from Heidelberg

(Germany) and Hyytiälä (Finland). The methods for correlation analysis were developed

further in this paper. The values for activation and kinetic coefficients in these

data sets were determined. I made half of the data analysis and participated in writing

the article.

Paper III studies the effect of relative humidity on atmospheric particle formation

by combining data analysis and aerosol dynamical modelling. Utilizing the QUEST

II data set, the paper presents an anticorrelation of sulphuric acid concentration and

particle formation rate with relative humidity. Aerosol dynamical modelling is used to

investigate thereasonsfortheobserved anticorrelation, concluding thatrelativehumidity

suppresses nucleation due to decreased production of sulphuric acid. I participated

in the data analysis and assisted in finalizing the manuscript.

Paper IV examines factors affecting the correlation of sulphuric acid and particle

formation rate with the aid of aerosol dynamical modelling. Especially, the effect of

the nucleation mechanism and condensational growth on the observed correlation at 3

nm are investigated. The paper demonstrates that the correlation with sulphuric acid

can change significantly during the particle growth from nucleation size to 3 nm. I

made most of the simulation runs and data processing, and was responsible for writing

the article.

Paper V investigates the methods used to calculate particle formation rates from the

time evolution of the particle size distribution. The study is based on the analysis of

a simulated particle formation event. The paper compares different particle formation

rate calculation methods and gives a recommendation of the most appropriate one. I

participated in planning the paper, gave ideas for the data analysis and assisted in

writing the manuscript.

Paper VI presents an analysis of one-year measurements of cloud condensation nuclei


ThecriticaldiametersforCCNactivationandthehygroscopicity parameter

kappa were derived from the data. The seasonal variation of CCN concentrations and

critical diameters were investigated, as well as the effect of new particle formation on

them. I made most of the analysis related to CCN concentration measurements and

was the main author of the paper. The hygroscopicity data was analysed by other



7 Summary and conclusions

Particle formation from gaseous precursors is a major source of new particles in the

atmosphere. Due to their ability to scatter solar radiation and influence cloud formation,

aerosol particles cause a net cooling effect on the Earth’s climate, thus partly

counteracting the climate warming caused by greenhouse gases. The understanding of

atmospheric aerosol formation, their growth to climatically relevant sizes and role in

cloud formation is crucial for reliable modelling of the climate system.

Atmospheric new particle formation or nucleation is a complex phenomenon to be

handled both experimentally and theoretically. Sulphuric acid (together with water

vapour) is assumed to be one of the key compounds in atmospheric nucleation, but the

exact nucleation mechanism and the identity of other nucleating compounds, such as

ammonia, amines or some oxygenated organic compounds, are still unknown.

At the start of the research made in this thesis, aerosol instrumentation was limited to

measure only particles larger than 3 nm in diameter. This detection limit prevented

researchers from performing direct measurements of atmospheric nucleation. One aim

of examining the correlation of particle formation with sulphuric acid, the assumed

main nucleating substance, was to obtain indirect information on the processes below

thedetectionlimit: forexample, fromthetimeshiftanalysisthegrowthrateofparticles

< 3 nm could be estimated. The gap between 3 nm and the nucleation size at 1–1.5 nm

was crossed using theoretical methods to account for particle losses in between those


The main part of the research performed in this thesis was devoted to the investigation

of the correlation of new particle formation with the sulphuric acid concentration

(papers I–IV). The number concentration of freshly nucleated particles (3–6 nm in

diameter) as well as the new particle formation rate were observed to correlate with

the sulphuric acid concentration to the power of 1–2. Based on this correlation, new

semi-empirical parameterisations for the atmospheric nucleation rate were developed:

activation nucleation with a linear dependence and kinetic nucleation with a squared

dependence onthesulphuric acidconcentration. Thevalues oftheproportionalitycoefficients

for these nucleation mechanisms, activation coefficient A and kinetic coefficient

K, were determined for three data sets from two different environments in Finland and

Germany. The determination of the empirical A and K coefficients can be considered

as the most valuable result of this thesis. The developed parameterisations, especially

the activation type nucleation, have been used widely by other reseachers in aerosol

dynamical and climate models.

ThelargevariationsinbothAandK (two andthreeordersofmagnitude, respectively),

even at the same measurement site, suggest that there are other important factors

affecting the nucleation process in addition to sulphuric acid. The parameterisations

have been further developed by Paasonen et al. (2010) to include the possible effect

of organic vapours. However, the exact physical and chemical details hidden in the

empirical A and K coefficients remain still unknown. More research, especially insitu

measurements of the chemical composition of nucleated clusters, is needed to


explain the large variation of nucleation coefficients and to develop more accurate

parameterisations for nucleation.

The connection between particle formation and sulphuric acid concentration was further

investigated by means of aerosol dynamical modelling, applying the activation and

kinetic nucleation mechanisms (paper IV). It was found that the correlation exponent

with sulphuric acid concentration was different for the formation rate of 1.5 nm

particles (J 1.5 i.e. the nucleation rate), for the formation rate of 3 nm particles (J 3 ),

and for the number concentration of 3–6 particles (N 3−6 ). When going from J 1.5 to J 3

and N 3−6 , the value of the correlation exponent could increase by 1–3 units, the change

depending on the particle growth rate and especially its profile as a function of particle

size. In order to obtain close to linear dependence for N 3−6 , activation nucleation and

a negligible saturation concentration for the condensable organics were required, i.e.

the growth rate of freshly nucleated particles needed to be high.

Atmospheric new particle formation happens preferably in conditions with low relative

humidity, which has been explained mainly by high condensation and coagulation

sinks due to water uptake by particles at high humidities. This thesis produced a

new possible hypothesis for the effect of relative humidity on new particle formation

(paper III). The particle formation rate seems to be limited by decreased production

of sulphuric acid at high relative humidities, due to decreased solar radiation reaching

the ground at humid conditions (where clouds may be present). However, a clear

causal relationship between the decreased solar radiation and relative humidity could

not be proven, because these variables have opposite diurnal profiles, thereby creating

an apparent anticorrelation between them. Aerosol dynamical modelling revealed that,

in comparison to the reduced sulphuric acid effect, the previously suggested effects of

increased condensation and coagulation sinks have a smaller contribution in inhibiting

nucleation at high humidities.

With regard to research methods, in this thesis the methods for analysing the correlation

of particle formation rate with sulphuric acid were developed (papers I, II and

IV). Additionally, the accuracy of the conventional method to estimate particle formation

rate from the particle size distribution data was evaluated and a new, slightly

improved method for J 3 calculation was proposed (paper V).

The importance of aerosols in the climate system is primarily associated with their

ability to act as cloud condensation nuclei. The potential of boreal forest aerosols to

act as cloud condensation nuclei was studied based on 1 year of CCN concentration

measurements and hygroscopicity measurements at ground level at the SMEAR II station

(paper VI). The CCN concentration was found to have a seasonal variation with

highest concentrations in summertime, when also the particle growth rates caused by

biogenic organic vapours are the highest. New particle formation events were observed

to almost double the CCN concentrations on the day following the event. Estimates for

the critical diameter of cloud droplet activation and for the hygroscopcity parameter κ

were determined. The values corresponded to a mixture of mainly organicaerosol, with

a small fractionof inorganic sulfates, as expected in the boreal forest environment. The

determined κ-values can be applied in modelling the CCN-activation of boreal forest


aerosols in climate models.

By the time this thesis was finalized, insummer-autumn 2013, huge advances inaerosol

measurement technology have been achieved. Several new developed instruments —

such as AIS, NAIS, PHA-CPC, PSM — have gone below the 3 nm limit, and opened

up the world of aerosol science directly down to the nucleation size at 1–2 nm. This

has increased our understanding of the complicated processes related to atmospheric

new particle formation. Mass spectrometric techniques and the particle size magnifier

(PSM) are closing the gap between molecular clusters (1 nm) and aerosol particles

(> 2–3 nm): in laboratory conditions the growth of nucleated clusters can be followed

molecule by molecule (Kirkby et al., 2011; Almeida et al., 2013). Still, many things

remain to be investigated, e.g.: What is the role of other substances than sulphuric

acid (amines, various organic molecules/vapours) in nucleation and early growth of

nucleated clusters? What is the molecular structure of nucleated and pre-nucleation

clusters? In theobserved cluster pool, arethere different kinds of clusters, and which of

them start to grow further? In the atmosphere, the variety of compounds (especially of

organiccompounds)issohuge, thatthereiscertainlyworkalsoforthenextgenerations.

Measurement data both from field and laboratory are needed to get real information

on the quantities and processes under study. In atmospheric data — all processes

happening under one common sun — there are many correlations, but fewer, and

often complex causal relationships between the variables. Careful data-analysis and

modelling can give insights on how the measurement data can be interpreted: what is

the importance of a certain possible reason behind an observed phenomenon, and what

can be concluded based on the data.

As an end product, research will provide useful tools, for example in the form of

nucleation rate parameterisations. These can be applied in regional aerosol and global

climate models, for making predictions of the future and suggestions for the actions

needed to prevent climate change or to improve air quality — for the good of society

and mankind.



Aalto, P., Hämeri, K., Becker, E., Weber, R., Salm, J., Mäkelä, J. M., Hoell, C., O’Dowd, C. D.,

Karlsson, H., Hansson, H., Väkevä, M., Koponen, I., Buzorius, G., and Kulmala, M.: Physical

characterization of aerosol particles during nucleation events, Tellus B, 53, 344–358, 2001.

Almeida, J., Schobesberger, S., Kürten, A., Ortega, I. K., Kupiainen-Määttä, O., Praplan, A. P.,

Adamov, A., Amorim, A., Bianchi, F., Breitenlechner, M., et al.: Molecular understanding of

sulphuric acid-amine particle nucleation in the atmosphere, Nature, 502, 359–363, 2013.

Andreae, M., Jones, C., and Cox, P.: Strong present-day aerosol cooling implies a hot future, Nature,

435, 1187–1190, 2005.

Anttila, P., Rissanen, T., Shimmo, M., Kallio, M., Hyötyläinen, T., Kulmala, M., and Riekkola, M.:

Organic compounds in atmospheric aerosols from a Finnish coniferous forest, Boreal Env. Res., 10,

371–384, 2005.

Anttila, T., Kerminen, V.-M., and Lehtinen, K. E. J.: Parameterizing the formation rate of new

particles: The effect of nuclei self-coagulation, J. Aerosol Sci., 41, 621–636, 2010.

Archer, D. and Pierrehumbert, R., eds.: The Warming papers: The Scientific Foundation for the

Climate Change Forecast, Wiley-Blackwell, John Wiley & Sons Ltd., United Kingdom, 2011.

Arneth, A., Unger, N., Kulmala, M., and Andreae, M. O.: Clean the Air, Heat the Planet?, Science,

326, 672–673, 2009.

Asmi, A., Coen, M. C., Ogren, J. A., Andrews, E., Sheridan, P., Jefferson, A., Weingartner, E.,

Baltensperger, U., Bukowiecki, N., Lihavainen, H., Kivekäs, N., Asmi, E., Aalto, P. P., Kulmala,

M., Wiedensohler, A., Birmili, W., Hamed, A., O’Dowd, C., Jennings, S. G., Weller, R., Flentje,

H., Fjaeraa, A. M., Fiebig, M., Myhre, C. L., Hallar, A. G., Swietlicki, E., Kristensson, A., and

Laj, P.: Aerosol decadal trends - Part 2: In-situ aerosol particle number concentrations at GAW

and ACTRIS stations, Atmos. Chem. Phys., 13, 895–916, 2013.

Asmi, E., Sipilä, M., Manninen, H. E., Vanhanen, J., Lehtipalo, K., Gagné, S., Neitola, K., Mirme,

A., Mirme, S., Tamm, E., Uin, J., Komsaare, K., Attoui, M., and Kulmala, M.: Results of the first

air ion spectrometer calibration and intercomparison workshop, Atmos. Chem. Phys., 9, 141–154,


Asmi, E., Kivekäs, N., Kerminen, V.-M., Komppula, M., Hyvärinen, A.-P., Hatakka, J., Viisanen, Y.,

and Lihavainen, H.: Secondary new particle formation in Northern Finland Pallas site between the

years 2000 and 2010, Atmos. Chem. Phys., 11, 12959–12972, 2011.

Aufmhoff, H., Hanke, M., Uecker, J., Schlager, H., and Arnold, F.: An ion trap CIMS instrument for

combined measurements of atmospheric OH and H 2 SO 4 : First test measurements above and inside

the planetary boundary layer, Int. J. Mass Spectrom., 308, 26–34, 2011.

Ayers, G., Gillett, R., and Gras, J.: On the vapor pressure of sulfuric acid, Geophys. Res. Lett., 7,

433–436, 1980.

Ball, S. M., Hanson, D. R., Eisele, F. L., and McMurry, P. H.: Laboratory studies of particle nucleation:

Initial results for H 2 SO 4 , H 2 O, and NH 3 vapors, J. Geophys. Res., 104, 23709–23718,


Bennartz, R., Fan, J., Rausch, J., Leung, L. R., and Heidinger, A. K.: Pollution from China increases

cloud droplet number, suppresses rain over the East China Sea, Geophys. Res. Lett., 38, 2011.

Benson, D. R., Young, L.-H., Kameel, F. R., and Lee, S.-H.: Laboratory-measured nucleation rates

of sulfuric acid and water binary homogeneous nucleation from the SO 2 + OH reaction, Geophys.

Res. Lett., 35, L11801, 2008.


Berndt, T., Böge, O., Stratmann, F., Heintzenberg, J., and Kulmala, M.: Rapid formation of sulfuric

acid particles at near-atmospheric conditions, Science, 307, 698–700, 2005.

Berndt, T., Stratmann, F., Brasel, S., Heintzenberg, J., Laaksonen, A., and Kulmala, M.: SO 2

oxidation products other than H 2 SO 4 as a trigger of new particle formation. Part 1: Laboratory

investigations, Atmos. Chem. Phys., 8, 6365–6374, 2008.

Berresheim, H., Elste, T., Plass-Dülmer, C., Eisele, F., and Tanner, D.: Chemical ionization mass

spectrometer for long-term measurements of atmospheric OH and H 2 SO 4 , Int. J. Mass Spectrom.,

202, 91–109, 2000.

Birmili, W., Berresheim, H., Plass-Dülmer, C., Elste, T., Gilge, S., Wiedensohler, A., and Uhrner, U.:

The Hohenpeissenberg aerosol formation experiment (HAFEX): a long-term study including sizeresolved

aerosol, H 2 SO 4 , OH, and monoterpenes measurements, Atmos. Chem. Phys., 3, 361–376,


Birmili, W., Wiedensohler, A., Plass-Dülmer, C., and Berresheim, H.: Evolution of newly formed

aerosol particles in the continental boundary layer: A case study including OH and H 2 SO 4 measurements,

Geophys. Res. Lett., 27, 2205–2208, 2000.

Boy, M. and Kulmala, M.: Nucleation events in the continental boundary layer: Influence of physical

and meteorological parameters, Atmos. Chem. Phys., 2, 1–16, 2002.

Boy, M., Rannik, U., Lehtinen, K. E. J., Tarvainen, V., Hakola, H., and Kulmala, M.: Nucleation

events in the continental boundary layer: Long-term statistical analyses of aerosol relevant characteristics,

J. Geophys. Res., 108, 4667, 2003.

Boy, M., Kulmala, M., Ruuskanen, T., Pihlatie, M., Reissell, A., Aalto, P. P., Keronen, P., Dal Maso,

M., Hellén, H., Hakola, H., Jansson, R., Hanke, M., and Arnold, F.: Sulphuric acid closure and

contribution to nucleation mode particle growth, Atmos. Chem. Phys., 5, 863–878, 2005.

Boy, M., Hellmuth, O., Korhonen, H., Nilsson, E., ReVelle, D., Turnipseed, A., Arnold, F., and

Kulmala, M.: MALTE - model to predict new aerosol formation in the lower troposphere, Atmos.

Chem. Phys., 6, 4499–4517, 2006.

Brunekreef, B. and Holgate, S. T.: Air pollution and health, Lancet, 360, 1233–1242, 2002.

Brus, D., Hyvärinen, A. P., Viisanen, Y., Kulmala, M., and Lihavainen, H.: Homogeneous nucleation

of sulfuric acid and water mixture: experimental setup and first results, Atmos. Chem. Phys., 10,

2631–2641, 2010.

Brus, D., Neitola, K., Hyvärinen, A. P., Petäjä, T., Vanhanen, J., Sipilä, M., Paasonen, P., Kulmala,

M., and Lihavainen, H.: Homogenous nucleation of sulfuric acid and water at close to atmospherically

relevant conditions, Atmos. Chem. Phys., 11, 5277–5287, 2011.

Carlton, A. G., Wiedinmyer, C., and Kroll, J. H.: A review of Secondary Organic Aerosol (SOA)

formation from isoprene, Atmos. Chem. Phys., 9, 4987–5005, 2009.

Chang, D., Song, Y., and Liu, B.: Visibility trends in six megacities in China 1973–2007,Atmos. Res.,

94, 161–167, 2009a.

Chang, L.-S., Schwartz, S. E., McGraw, R., and Lewis, E. R.: Sensitivity of aerosol properties to new

particle formation mechanism and to primary emissions in a continental-scale chemical transport

model, J. Geophys. Res., 114, D07203, 2009b.

Dal Maso, M., Kulmala, M., Lehtinen, K. E. J., Mäkelä, J., Aalto, P., and O’Dowd, C. D.: Condensation

and coagulation sinks and formation of nucleation mode particles in coastal and boreal forest

boundary layers, J. Geophys. Res., 107, 8097, 2002.


Dal Maso, M., Kulmala, M., Riipinen, I., Wagner, R., Hussein, T., Aalto, P. P., and Lehtinen, K.

E. J.: Formation and growth of fresh atmospheric aerosols: Eight years of aerosol size distribution

data from SMEAR II, Hyytiälä, Finland, Boreal Env. Res., 10, 323–336, 2005.

Deshler, T.: A review of global stratospheric aerosol: Measurements, importance, life cycle, and local

stratospheric aerosol, Atmos. Res., 90, 223–232, 2008.

Dockery, D. and Pope, C.: Acute respiratory effects of particulate air-pollution, Annu. Rev. Public

Health, 15, 107–132, 1994.

Doyle, G.: Self-Nucleation in the Sulfuric Acid-Water System, J. Chem. Phys., 35, 795, 1961.

Dusek, U., Frank, G., Hildebrandt, L., Curtius, J., Schneider, J., Walter, S., Chand, D., Drewnick,

F., Hings, S., Jung, D., Borrmann, S., and Andreae, M.: Size matters more than chemistry for

cloud-nucleating ability of aerosol particles, Science, 312, 1375–1378, 2006.

Dusek, U., Frank, G. P., Curtius, J., Drewnick, F., Schneider, J., Kürten, A., Rose, D., Andreae,

M. O., Borrmann, S., and Pöschl, U.: Enhanced organicmassfraction and decreasedhygroscopicity

of cloud condensation nuclei (CCN) during new particle formation events, Geophys. Res. Lett., 37,


Ehn, M., Petäjä, T., Aufmhoff, H., Aalto, P., Hameri, K., Arnold, F., Laaksonen, A., and Kulmala,

M.: Hygroscopic properties of ultrafine aerosol particles in the boreal forest: diurnal variation,

solubility and the influence of sulfuric acid, Atmos. Chem. Phys., 7, 211–222, 2007.

Ehn, M., Vuollekoski, H., Petäjä, T., Kerminen, V.-M., Vana, M., Aalto, P., de Leeuw, G., Ceburnis,

D., Dupuy, R., O’Dowd, C. D., and Kulmala, M.: Growth rates during coastal and marine new

particle formation in western Ireland, J. Geophys. Res., 115, D18218, 2010.

Engelstaedter, S., Tegen, I., and Washington, R.: North African dust emissions and transport, Earth-

Sci. Reviews, 79, 73–100, 2006.

Fiedler, V., Dal Maso, M., Boy, M., Aufmhoff, H., Hoffmann, J., Schuck, T., Birmili, W., Hanke,

M., Uecker, J., Arnold, F., and Kulmala, M.: The contribution of sulphuric acid to atmospheric

particle formation and growth: a comparison between boundary layers in Northern and Central

Europe, Atmos. Chem. Phys., 5, 1773–1785, 2005.

Gagné, S., Nieminen, T., Kurtén, T., Manninen, H. E., Petäjä, T., Laakso, L., Kerminen, V.-M., Boy,

M., and Kulmala, M.: Factors influencing the contribution of ion-induced nucleation in a boreal

forest, Finland, Atmos. Chem. Phys., 10, 3743–3757, 2010.

Gagné, S., Leppä, J., Petäjä, T., McGrath, M. J., Vana, M., Kerminen, V., Laakso, L., and Kulmala,

M.: Aerosol charging state at an urban site: new analytical approach and implications for ioninduced

nucleation, Atmos. Chem. Phys., 12, 4647–4666, 2012.

Gao, J., Chai, F., Wang, T., and Wang, W.: Particle number size distribution and new particle

formation (NPF) in Lanzhou, Western China, Particuology, 9, 611–618, 2011.

Gaydos, T. M., Stanier, C. O., and Pandis, S. N.: Modeling of in situ ultrafine atmospheric particle

formation in the eastern United States, J. Geophys. Res., 110, D07S12, 2005.

Grini, A., Korhonen, H., Lehtinen, K. E. J., Isaksen, I., and Kulmala, M.: A combined photochemistry/aerosol

dynamics model: model development and a study of new particle formation, Boreal

Env. Res., 10, 525–541, 2005.

Gunthe, S. S., King, S. M., Rose, D., Chen, Q., Roldin, P., Farmer, D. K., Jimenez, J. L., Artaxo,

P., Andreae, M. O., Martin, S. T., and Pöschl, U.: Cloud condensation nuclei in pristine tropical

rainforest air of Amazonia: size-resolved measurements and modeling of atmospheric aerosol

composition and CCN activity, Atmos. Chem. Phys., 9, 7551–7575, 2009.


Haapanala, S., Rinne, J., Hakola, H., Hellén, H., Laakso, L., Lihavainen, H., Janson, R., O’Dowd, C.,

and Kulmala, M.: Boundary layer concentrations and landscape scale emissions of volatile organic

compounds in early spring, Atmos. Chem. Phys., 7, 1869–1878, 2007.

Hamed, A., Birmili, W., Joutsensaari, J., Mikkonen, S., Asmi, A., Wehner, B., Spindler, G., Jaatinen,

A., Wiedensohler, A., Korhonen, H., Lehtinen, K. E. J., and Laaksonen, A.: Changes in the

productionrateofsecondaryaerosolparticlesin CentralEuropeinviewofdecreasingSO 2 emissions

between 1996 and 2006, Atmos. Chem. Phys., 10, 1071–1091, 2010.

Hanke, M., Uecker, J., Reiner, T., and Arnold, F.: Atmospheric peroxy radicals: ROXMAS, a new

mass-spectrometric methodology for speciated measurements of HO 2 and ΣRO 2 and first results,

Int. J. Mass Spectrom., 213, 91–99, 2002.

Hari, P. and Kulmala, M.: Station for measuring ecosystem-atmosphere relations (SMEAR II), Boreal

Env. Res., 10, 315–322, 2005.

Haywood,J.andBoucher,O.: Estimatesofthedirectandindirectradiativeforcingduetotropospheric

aerosols: A review, Reviews of Geophysics, 38, 513–543, 2000.

Hirsikko, A., Laakso, L., Hõrrak, U., Aalto, P. P., Kerminen, V.-M., and Kulmala, M.: Annual and

size dependent variation of growth rates and ion concentrations in boreal forest, Boreal Env. Res.,

10, 357–369, 2005.

Hirsikko,A., Nieminen, T., Gagné,S., Lehtipalo,K.,Manninen,H.E.,Ehn, M.,Hõrrak,U., Kerminen,

V.-M., Laakso, L., McMurry, P. H., Mirme, A., Mirme, S., Petäjä, T., Tammet, H., Vakkari, V.,

Vana, M., and Kulmala, M.: Atmospheric ions and nucleation: a review of observations, Atmos.

Chem. Phys., 11, 767–798, 2011.

Hoppel, W., Frick, G., and Fitzgerald, J.: Deducing droplet concentration and supersaturation in

marine boundary layer clouds from surface aerosol measurements, J. Geophys. Res., 101, 26553–

26565, 1996.

Hultin, K. A. H., Krejci, R., Pinhassi, J., Gomez-Consarnau, L., Martensson, E. M., Hagstrom, A.,

and Nilsson, E. D.: Aerosol and bacterial emissions from Baltic Seawater, Atmos. Res., 99, 1–14,


Hussein, T., Hämeri, K., Aalto, P., Asmi, A., Kakko, L., and Kulmala, M.: Particle size characterization

and the indoor-to-outdoor relationship of atmospheric aerosols in Helsinki, Scand. J. Work

Environ. Health, 30, 54–62, 2004.

Hyvönen, S., Junninen, H., Laakso, L., Dal Maso, M., Grönholm, T., Bonn, B., Keronen, P., Aalto,

P., Hiltunen, V., Pohja, T., Launiainen, S., Hari, P., Mannila, H., and Kulmala, M.: A look at

aerosol formation using data mining techniques, Atmos. Chem. Phys., 5, 3345–3356, 2005.

IPCC: Summary for Policymakers. In: Climate Change 2007: The Physical Science Basis. Contribution

of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on

Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor

and H.L. Miller (eds.)], Cambridge University Press, Cambridge, United Kingdom and New York,

NY, USA, 2007.

Jefferson, A., Eisele, F., Ziemann, P., Weber, R., Marti, J., and McMurry, P.: Measurements of the

H 2 SO 4 mass accommodation coefficient onto polydisperse aerosol, J. Geophys. Res., 102, 19021–

19028, 1997.

Jimenez, J. L., Canagaratna, M. R., Donahue, N. M., Prevot, A. S. H., Zhang, Q., Kroll, J. H.,

DeCarlo, P. F., Allan, J. D., Coe, H., Ng, N. L., Aiken, A. C., Docherty, K. S., Ulbrich, I. M.,

Grieshop, A. P., Robinson, A. L., Duplissy, J., Smith, J. D., Wilson, K. R., Lanz, V. A., Hueglin,

C., Sun, Y. L., Tian, J., Laaksonen, A., Raatikainen, T., Rautiainen, J., Vaattovaara, P., Ehn,


M., Kulmala, M., Tomlinson, J. M., Collins, D. R., Cubison, M. J., Dunlea, E. J., Huffman, J. A.,

Onasch, T. B., Alfarra, M. R., Williams, P. I., Bower, K., Kondo, Y., Schneider, J., Drewnick, F.,

Borrmann,S., Weimer, S., Demerjian, K., Salcedo, D., Cottrell, L., Griffin, R., Takami, A., Miyoshi,

T., Hatakeyama, S., Shimono, A., Sun, J. Y., Zhang, Y. M., Dzepina, K., Kimmel, J. R., Sueper,

D., Jayne, J. T., Herndon, S. C., Trimborn, A. M., Williams, L. R., Wood, E. C., Middlebrook,

A. M., Kolb, C. E., Baltensperger, U., and Worsnop, D. R.: Evolution of Organic Aerosols in the

Atmosphere, Science, 326, 1525–1529, 2009.

Jung, J. G., Pandis, S. N., and Adams, P. J.: Evaluation of nucleation theories in a sulfur-rich

environment, Aerosol Sci. Tech., 42, 495–504, 2008.

Kerminen, V.-M. and Kulmala, M.: Analytical formulae connecting the ”real” and the ”apparent”

nucleation rate and the nuclei number concentration for atmospheric nucleation events, J. Aerosol

Sci., 33, 609–622, 2002.

Kerminen, V.-M., Anttila, T., Lehtinen, K. E. J., and Kulmala, M.: Parameterizationfor atmospheric

new-particle formation: Application to a system involving sulfuric acid and condensable watersoluble

organic vapors, Aerosol Sci. Tech., 38, 1001–1008, 2004.

Kerminen, V.-M., Paramonov, M., Anttila, T., Riipinen, I., Fountoukis, C., Korhonen, H., Asmi, E.,

Laakso, L., Lihavainen, H., Swietlicki, E., Svenningsson, B., Asmi, A., Pandis, S. N., Kulmala, M.,

and Petäjä, T.: Cloud condensation nuclei production associated with atmospheric nucleation: a

synthesis based on existing literature and new results, Atmos. Chem. Phys, 12, 12037–12059, 2012.

Kirkby, J., Curtius, J., Almeida, J., Dunne, E., Duplissy, J., Ehrhart, S., Franchin, A., Gagné,

S., Ickes, L., Kuerten, A., Kupc, A., Metzger, A., Riccobono, F., Rondo, L., Schobesberger, S.,

Tsagkogeorgas,G., Wimmer, D., Amorim, A., Bianchi, F., Breitenlechner, M., David, A., Dommen,

J., Downard, A., Ehn, M., Flagan, R. C., Haider, S., Hansel, A., Hauser, D., Jud, W., Junninen,

H., Kreissl, F., Kvashin, A., Laaksonen, A., Lehtipalo, K., Lima, J., Lovejoy, E. R., Makhmutov,

V., Mathot, S., Mikkilä, J., Minginette, P., Mogo, S., Nieminen, T., Onnela, A., Pereira, P., Petäjä,

T., Schnitzhofer, R., Seinfeld, J. H., Sipilä, M., Stozhkov, Y., Stratmann, F., Tome, A., Vanhanen,

J., Viisanen, Y., Vrtala, A., Wagner, P. E., Walther, H., Weingartner, E., Wex, H., Winkler, P. M.,

Carslaw, K. S., Worsnop, D. R., Baltensperger, U., and Kulmala, M.: Role of sulphuric acid,

ammonia and galactic cosmic rays in atmospheric aerosol nucleation, Nature, 476, 429–433, 2011.

Koch, D. and Del Genio, A. D.: Black carbon semi-direct effects on cloud cover: review and synthesis,

Atmos. Chem. Phys., 10, 7685–7696, 2010.

Koehler, K. A., Kreidenweis, S. M., DeMott, P. J., Petters, M. D., Prenni, A. J., and Carrico, C. M.:

Hygroscopicity and cloud droplet activation of mineral dust aerosol, Geophys. Res. Lett., 36, 2009.

Komppula, M., Sihto, S.-L., Korhonen, H., Lihavainen, H., Kerminen, V.-M., Kulmala, M., and

Viisanen, Y.: New particle formation in air mass transported between two measurement sites in

Northern Finland, Atmos. Chem. Phys., 6, 2811–2824, 2006.

Koponen, I., Virkkula, A., Hillamo, R., Kerminen, V.-M., and Kulmala, M.: Number sizedistributions

and concentrations of marine aerosols: Observations during a cruise between the English Channel

and the coast of Antarctica, J. Geophys. Res., 107, 2002.

Koponen, I., Virkkula, A., Hillamo, R., Kerminen, V.-M., and Kulmala, M.: Number size distributions

and concentrations of the continental summer aerosols in Queen Maud Land, Antarctica, J.

Geophys. Res., 108, 2003.

Korhonen, H., Lehtinen, K. E. J., and Kulmala, M.: Multicomponent aerosoldynamics model UHMA:

model development and validation, Atmos. Chem. Phys., 4, 757–771, 2004.


Korhonen, H., Kerminen, V.-M., Lehtinen, K. E. J., and Kulmala, M.: CCN activation and cloud

processing in sectional aerosol models with low size resolution, Atmos. Chem. Phys., 5, 2561–2570,


Korhonen, P., Kulmala, M., Laaksonen, A., Viisanen, Y., McGraw, R., and Seinfeld, J.: Ternary

nucleation of H2SO4, NH3, and H2O in the atmosphere, J. Geophys. Res., 104, 26349–26353,


Kuang, C., McMurry, P., McCormick, A., and Eisele, F.: Dependence of nucleation rates on sulfuric

acid vapor concentration in diverse atmospheric locations, J. Geophys. Res., 113, D10209, 2008.

Kulmala, M.: How particles nucleate and grow, Science, 302, 1000–1001, 2003.

Kulmala, M. and Kerminen, V.-M.: On the formation and growth of atmospheric nanoparticles,

Atmos. Res., 90, 132–150, 2008.

Kulmala, M., Laaksonen, A., and Pirjola, L.: Parameterizations for sulfuric acid/water nucleation

rates, J. Geophys. Res., 103, 8301–8307, 1998a.

Kulmala, M., Toivonen, A., Mäkelä, J., and Laaksonen, A.: Analysis of the growth of nucleation mode

particles observed in Boreal forest, Tellus B, 50, 449–462, 1998b.

Kulmala, M., Dal Maso, M., Mäkelä, J., Pirjola, L., Väkevä, M., Aalto, P., Miikkulainen, P., Hameri,

K., and O’Dowd, C. D.: On the formation, growth and composition of nucleation mode particles,

Tellus B, 53, 479–490, 2001a.

Kulmala, M., Hameri, K., Aalto, P. P., Mäkelä, J., Pirjola, L., Nilsson, E., Buzorius, G., Rannik, U.,

Dal Maso, M., Seidl, W., Hoffman, T., Janson, R., Hansson, H., Viisanen, Y., Laaksonen, A., and

O’Dowd, C. D.: Overview of the international project on biogenic aerosol formation in the boreal

forest (BIOFOR), Tellus B, 53, 324–343, 2001b.

Kulmala, M., Boy, M., Suni, T., Gaman, A., Raivonen, M., Aaltonen, V., Adler, H., Anttila, T.,

Fiedler, V., Gronholm, T., Hellén, H., Herrmann, E., Jalonen, R., Jussila, M., Komppula, M.,

Kosmale, M., Plauskaite, K., Reis, R., Savola, N., Soini, P., Virtanen, S., Aalto, P., Dal Maso, M.,

Hakola, H., Keronen, P., Vehkamäki, H., Rannik, U., Lehtinen, K. E. J., and Hari, P.: Aerosols in

boreal forest: wintertime relations between formation events and bio-geo-chemical activity, Boreal

Env. Res., 9, 63–74, 2004a.

Kulmala, M., Kerminen, V.-M., Anttila, T., Laaksonen, A., and O’Dowd, C. D.: Organic aerosol

formation via sulphate cluster activation, J. Geophys. Res., 109, D04205, 2004b.

Kulmala, M., Suni, T., Lehtinen, K. E. J., Dal Maso, M., Boy, M., Reissell, A., Rannik, U., Aalto,

P., Keronen, P., Hakola, H., Back, J., Hoffmann, T., Vesala, T., and Hari, P.: A new feedback

mechanism linking forests, aerosols, and climate, Atmos. Chem. Phys., 4, 557–562, 2004c.

Kulmala, M., Vehkamäki, H., Petäjä, T., Dal Maso, M., Lauri, A., Kerminen, V.-M., Birmili, W.,

and McMurry, P.: Formation and growth rates of ultrafine atmospheric particles: a review of

observations, J. Aerosol Sci., 35, 143–176, 2004d.

Kulmala, M., Lehtinen, K. E. J., and Laaksonen, A.: Cluster activation theory as an explanation of

the linear dependence between formation rate of 3 nm particles and sulphuric acid concentration,

Atmos. Chem. Phys., 6, 787–793, 2006.

Kulmala, M., Riipinen, I., Sipilä, M., Manninen, H. E., Petäjä, T., Junninen, H., Dal Maso, M.,

Mordas, G., Mirme, A., Vana, M., Hirsikko, A., Laakso, L., Harrison, R., Hanson, I., Leung, C.,

Lehtinen, K. E. J., and Kerminen, V.-M.: Toward direct measurement of atmospheric nucleation,

Science, 318, 89–92, 2007.


Kulmala, M., Kontkanen, J., Junninen, H., Lehtipalo, K., Manninen, H. E., Nieminen, T., Petäjä,

T., Sipilä, M., Schobesberger, S., Rantala, P., Franchin, A., Jokinen, T., Järvinen, E., Äijälä,

M., Kangasluoma, J., Hakala, J., Aalto, P. P., Paasonen, P., Mikkilä, J., Vanhanen, J., Aalto, J.,

Hakola, H., Makkonen, U., Ruuskanen, T., Mauldin, R. L., Duplissy, J., Vehkamäki, H., Bäck, J.,

Kortelainen, A., Riipinen, I., Kurtén, T., Johnston, M. V., Smith, J. N., Ehn, M., Mentel, T. F.,

Lehtinen, K. E. J., Laaksonen, A., Kerminen, V.-M., and Worsnop, D. R.: Direct Observations of

Atmospheric Aerosol Nucleation, Science, 339, 943–946, 2013.

Kulmala, M., Petäjä, T., Nieminen, T., Sipilä, M., Manninen, H. E., Lehtipalo, K., Dal Maso, M.,

Aalto, P. P., Junninen, H., Paasonen, P., Riipinen, I., Lehtinen, K. E. J., Laaksonen, A., and

Kerminen, V.-M.: Measurementofthenucleationofatmosphericaerosolparticles,Nature Protocols,

7, 1651–1667, 2012.

Kurtén, T., Torpo, L., Ding, C., Vehkamäki, H., Sundberg, M., Laasonen, K., and Kulmala, M.: A

density functional study on water-sulfuric acid-ammonia clusters and implications for atmospheric

cluster formation, J. Geophys. Res., 112, 2007.

Kurtén, T., Loukonen, V., Vehkamäki, H., and Kulmala, M.: Amines are likely to enhance neutral

and ion-induced sulfuric acid-water nucleation in the atmosphere more effectively than ammonia,

Atmos. Chem. Phys., 8, 4095–4103, 2008.

Kuwata, M., Kondo, Y., Miyazaki, Y., Komazaki, Y., Kim, J. H., Yum, S. S., Tanimoto, H., and

Matsueda, H.: Cloud condensation nuclei activity at Jeju Island, Korea in spring 2005, Atmos.

Chem. Phys., 8, 2933–2948, 2008.

Kyrö, E. M., Kerminen, V. M., Virkkula, A., Dal Maso, M., Parshintsev, J., Ruiz-Jimenez, J.,

Forsstrom, L., Manninen, H. E., Riekkola, M. L., Heinonen, P., and Kulmala, M.: Antarctic new

particle formation from continental biogenic precursors, Atmos. Chem. Phys., 13, 3527–3546, 2013.

Laakso, L., Grönholm, T., Rannik, Ü., Kosmale, M., Fiedler, V., Vehkamäki, H., and Kulmala,

M.: Ultrafine particle scavenging coefficients calculated from 6 years field measurements, Atmos.

Environ., 37, 3605–3613, 2003.

Laakso, L., Petäjä, T., Lehtinen, K. E. J., Kulmala, M., Paatero, J., Hõrrak, U., Tammet, H.,

and Joutsensaari, J.: Ion production rate in a boreal forest based on ion, particle and radiation

measurements, Atmos. Chem. Phys., 4, 1933–1943, 2004.

Laaksonen, A., Kulmala, M., Berndt, T., Stratmann, F., Mikkonen, S., Ruuskanen, A., Lehtinen,

K. E. J., Dal Maso, M., Aalto, P., Petäjä, T., Riipinen, I., Sihto, S.-L., Janson, R., Arnold, F.,

Hanke, M., Ucker, J., Umann, B., Sellegri, K., O’Dowd, C. D., and Viisanen, Y.: SO 2 oxidation

products other than H 2 SO 4 as a trigger of new particle formation. Part 2: Comparison of ambient

and laboratory measurements, and atmospheric implications, Atmos. Chem. Phys., 8, 7255–7264,


Laitinen, T., Ehn, M., Junninen, H., Ruiz-Jimenez, J., Parshintsev, J., Hartonen, K., Riekkola, M.-L.,

Worsnop, D. R., and Kulmala, M.: Characterization of organic compounds in 10- to 50-nm aerosol

particles in borealforest with laserdesorption-ionizationaerosolmassspectrometer and comparison

with other techniques, Atmos. Environ., 45, 3711–3719, 2011.

Lauros, J., Nilsson, E., Dal Maso, M., and Kulmala, M.: Contribution of mixing in the ABL to new

particle formation based on observations, Atmos. Chem. Phys., 7, 4781–4792, 2007.

Lauros, J., Sogachev, A., Smolander, S., Vuollekoski, H., Sihto, S.-L., Mammarella, I., Laakso, L.,

Rannik, Ü., and Boy, M.: Particle concentration and flux dynamics in the atmospheric boundary

layer as the indicator of formation mechanism, Atmos. Chem. Phys., 11, 5591–5601, 2011.

Lehtinen, K. E. J. and Kulmala, M.: A model for particle formation and growth in the atmosphere

with molecular resolution in size, Atmos. Chem. Phys., 3, 251–257, 2003.


Lehtinen, K. E. J., Dal Maso, M., Kulmala, M., and Kerminen, V.-M.: Estimating nucleation rates

from apparent particle formation rates and vice versa: Revised formulation of the Kerminen-

Kulmala equation, J. Aerosol Sci., 38, 988–994, 2007.

Leppä, J., Kerminen, V.-M., Laakso, L., Korhonen, H., Lehtinen, K. E. J., Gagné, S., Manninen,

H. E., Nieminen, T., and Kulmala, M.: Ion-UHMA: a model for simulating the dynamics of neutral

and charged aerosol particles, Boreal Env. Res., 14, 559–575, 2009.

Leppä, J., Anttila, T., Kerminen, V. M., Kulmala, M., and Lehtinen, K. E. J.: Atmospheric new

particle formation: real and apparent growth of neutral and charged particles, Atmos. Chem.

Phys., 11, 4939–4955, 2011.

Lihavainen, H., Kerminen, V.-M., Komppula, M., Hatakka, J., Aaltonen, V., Kulmala, M., and

Viisanen, Y.: Production of ”potential” cloud condensation nuclei associated with atmospheric

new-particle formation in northern Finland, J. Geophys. Res., 108(D24), 4782, 2003.

Lihavainen, H., Kerminen, V.-M., Tunved, P., Aaltonen, V., Arola, A., Hatakka, J., Hyvärinen, A.,

and Viisanen, Y.: Observational signature of the direct radiative effect by natural boreal forest

aerosols and its relation to the corresponding first indirect effect, J. Geophys. Res., 114, D20206,


Lohmann, U. and Feichter, J.: Global indirect aerosol effects: a review, Atmos. Chem. Phys., 5,

715–737, 2005.

Lyubovtseva, Y., Sogacheva, L., Dal Maso, M., Bonn, B., Keronen, P., and Kulmala, M.: Seasonal

variations of trace gases, meteorological parameters, and formation of aerosols in boreal forests,

Boreal Env. Res., 10, 493–510, 2005.

Mäkelä, J., Aalto, P., Jokinen, V., Pohja, T., Nissinen, A., Palmroth, S., Markkanen, T., Seitsonen,

K., Lihavainen, H., and Kulmala, M.: Observations of ultrafine aerosol particle formation and

growth in boreal forest, Geophys. Res. Lett., 24, 1219–1222, 1997.

Makkonen, R., Asmi, A., Korhonen, H., Kokkola, H., Jarvenoja, S., Raisanen, P., Lehtinen, K.,

Laaksonen, A., Kerminen, V., Jarvinen, H., Lohmann, U., Bennartz, R., Feichter, J., and Kulmala,

M.: Sensitivity of aerosol concentrations and cloud properties to nucleation and secondary organic

distribution in ECHAM5-HAM global circulation model, Atmos. Chem. Phys., 9, 1747–1766, 2009.

Makkonen, R., Asmi, A., Kerminen, V.-M., Boy, M., Arneth, A., Hari, P., and Kulmala, M.: Air

pollution control and decreasing new particle formation lead to strong climate warming, Atmos.

Chem. Phys., 12, 1515–1524, 2012.

Malila, J.: Repairing the First Nucleation Theorem: Precritical Cluster Losses, Abstracts of the 19 th

International Conference on Nucleation and Atmospheric Aerosols, 2013.

Mammarella, I., Rannik, Ü., Aalto, P., Keronen, P., Vesala, T., and Kulmala, M.: Long-term aerosol

particleflux observations.PartII: Particlesize statisticsand deposition velocities, Atmos. Environ.,

45, 3794–3805, 2011.

Manninen, H., Nieminen, T., Riipinen, I., Yli-Juuti, T., Gagné, S., Asmi, E., Aalto, P., Petäjä, T.,

Kerminen, V., and Kulmala, M.: Charged and total particle formation and growth rates during

EUCAARI 2007 campaign in Hyytiälä, Atmos. Chem. Phys., 9, 4077–4089, 2009.

Manninen, H. E., Nieminen, T., Asmi, E., Gagné, S., Häkkinen, S., Lehtipalo, K., Aalto, P., Vana,

M., Mirme, A., Mirme, S., Hõrrak, U., Plass-Dülmer, C., Stange, G., Kiss, G., Hoffer, A., Toeroe,

N., Moerman, M., Henzing, B., de Leeuw, G., Brinkenberg, M., Kouvarakis, G., Bougiatioti, A.,

Mihalopoulos, N., O’Dowd, C., Ceburnis, D., Arneth, A., Svenningsson, B., Swietlicki, E., Tarozzi,

L., Decesari, S., Facchini, M., Birmili, W., Sonntag, A., Wiedensohler, A., Boulon, J., Sellegri,

K., Laj, P., Gysel, M., Bukowiecki, N., Weingartner, E., Wehrle, G., Laaksonen, A., Hamed, A.,


Joutsensaari, J., Petäjä, T., Kerminen, V.-M., and Kulmala, M.: EUCAARI ion spectrometer

measurements at 12 European sites - analysis of new particle formation events, Atmos. Chem.

Phys., 10, 7907–7927, 2010.

Mårtensson, E. M., Tunved, P., Korhonen, H., and Nilsson, E. D.: The role of sea-salt emissions

in controlling the marine Aitken and accumulation mode aerosol: a model study, Tellus B, 62,

259–279, 2010.

Marti, J., Jefferson, A., Cai, X., Richert, C., McMurry, P., and Eisele, F.: H 2 SO 4 vapor pressure of

sulfuric acid and ammonium sulfate solutions, J. Geophys. Res., 102(D3), 3725–3735, 1997.

Mauldin III, R., Berndt, T., Sipilä, M., Paasonen, P., Petäjä, T., Kim, S., Kurtén, T., Stratmann, F.,

Kerminen, V.-M., and Kulmala, M.: A new atmospherically relevant oxidant of sulphur dioxide,

Nature, 488, 193–196, 2012.

McFiggans, G., Artaxo, P., Baltensperger, U., Coe, H., Facchini, M. C., Feingold, G., Fuzzi, S., Gysel,

M., Laaksonen, A., Lohmann, U., Mentel, T. F., Murphy, D. M., O’Dowd, C. D., Snider, J. R.,

and Weingartner, E.: The effect of physical and chemical aerosol properties on warm cloud droplet

activation, Atmos. Chem. Phys., 6, 2593–2649, 2006.

McMurry, P. and Friedlander, S.: Aerosol formation in reacting gases - relation of surface-area to rate

of gas-to-particle conversion, J. Colloid Interface Sci., 64, 248–257, 1978.

McMurry, P. and Friedlander, S.: New particle formation in the presence of an aerosol, Atmospheric

Environment, 13, 1635–1651, 1979.

McMurry, P. and Friedlander, S.: Aerosol formation in reacting gases - relationship of surface-area to

rate of gas-to-particle conversion - Reply, J. Colloid Interface Sci., 74, 307–307, 1980.

McMurry, P. H., Fink, M., Sakurai, H., Stolzenburg, M. R., Mauldin, R. L., Smith, J., Eisele, F.,

Moore, K., Sjostedt, S., Tanner, D., Huey, L. G., Nowak, J. B., Edgerton, E., and Voisin, D.: A

criterion for new particle formation in the sulfur-rich Atlanta atmosphere, J. Geophys. Res., 110,

D22S02, 2005.

Merikanto, J., Napari, I., Vehkamäki, H., Anttila, T., and Kulmala, M.: New parameterization of

sulfuric acid-ammonia-water ternary nucleation rates at tropospheric conditions, J. Geophys. Res.,

112, 2007.

Merikanto, J., Spracklen, D. V., Mann, G. W., Pickering, S. J., and Carslaw, K. S.: Impact of

nucleation on global CCN, Atmos. Chem. Phys., 9, 8601–8616, 2009.

Metzger, A., Verheggen, B., Dommen, J., Duplissy, J., Prevot, A., Weingartner, E., Riipinen, I.,

Kulmala, M., Spracklen, D., Carslaw, K., and Baltensperger, U.: Evidence for the role of organics

in aerosol particle formation under atmospheric conditions, PNAS, 107, 6646–6651, 2010.

Mikkonen, S., Romakkaniemi, S., Smith, J. N., Korhonen, H., Petäjä, T., Plass-Dülmer, C., Boy, M.,

McMurry, P. H., Lehtinen, K. E. J., Joutsensaari, J., Hamed, A., Mauldin, I. R. L., Birmili, W.,

Spindler, G., Arnold, F., Kulmala, M., and Laaksonen, A.: A statistical proxy for sulphuric acid

concentration, Atmos. Chem. Phys., 11, 11319–11334, 2011.

Mitchell, J., Johns, T., Gregory, J., and Tett, S.: Climate response to increasing levels of greenhouse

gases and sulfate aerosols, Nature, 376, 501–504, 1995.

Mönkkönen, P., Koponen, I., Lehtinen, K. E. J., Hämeri, K., Uma, R., and Kulmala, M.: Measurements

in a highly polluted Asian mega city: observationsof aerosolnumber size distribution, modal

parameters and nucleation events, Atmos. Chem. Phys., 5, 57–66, 2005.

Napari, I., Kulmala, M., and Vehkamäki, H.: Ternary nucleation of inorganic acids, ammonia, and

water, J. Chem. Phys., 117, 8418–8425, 2002a.


Napari, I., Noppel, M., Vehkamäki, H., and Kulmala, M.: Parametrization of ternary nucleation rates

for H 2 SO 4 -NH 3 -H 2 O vapors, J. Geophys. Res., 107, 2002b.

Napari, I., Makkonen, R., Kulmala, M., and Vehkamäki, H.: Parameterization of ammonia and water

content of atmospheric droplets with fixed number of sulfuric acid molecules, Atmos. Res., 82,

514–522, 2006.

Nieminen, T., Manninen, H., Sihto, S.-L., Yli-Juuti, T., Mauldin, R. L., Petäjä, T., Riipinen, I.,

Kerminen, V.-M., and Kulmala, M.: Connection of sulfuric acid to atmospheric nucleation in

boreal forest, Environ. Sci. Technol., 43, 4715–4721, 2009.

Nilsson, E., Rannik, U., Kulmala, M., Buzorius, G., and O’Dowd, C. D.: Effects of continental

boundary layer evolution, convection, turbulence and entrainment, on aerosol formation, Tellus B,

53, 441–461, 2001.

O’Dowd, C. D. and Hoffmann, T.: Coastal new particle formation: A review of the current state-ofthe-art,

Environ. Chem., 2, 245–255, 2005.

O’Dowd, C. D., Jimenez, J., Bahreini, R., Flagan, R., Seinfeld, J., Hameri, K., Pirjola, L., Kulmala,

M., Jennings, S., and Hoffmann, T.: Marine aerosol formation from biogenic iodine emissions,

Nature, 417, 632–636, 2002.

Ortega, I. K., Kupiainen, O., Kurtén, T., Olenius, T., Wilkman, O., McGrath, M. J., Loukonen, V.,

and Vehkamäki, H.: From quantum chemical formation free energies to evaporation rates, Atmos.

Chem. Phys., 12, 225–235, 2012.

Paasonen, P., Sihto, S.-L., Nieminen, T., Vuollekoski, H., Riipinen, I., Plass-Dülmer, C., Berresheim,

H., Birmili, W., and Kulmala, M.: Connection between new particle formation and sulphuric acid

at Hohenpeissenberg (Germany) including the influence of organic compounds, Boreal Env. Res.,

14, 616–629, 2009.

Paasonen, P., Olenius, T., Kupiainen, O., Kurtén, T., Petäjä, T., Birmili, W., Hamed, A., Hu, M.,

Huey, L. G., Plass-Dülmer, C., Smith, J. N., Wiedensohler, A., Loukonen, V., McGrath, M. J.,

Ortega,I.K., Laaksonen,A., Vehkamäki, H., Kerminen, V.-M., andKulmala,M.: Ontheformation


new particle formation, Atmos. Chem. Phys., 12, 9113–9133, 2012.

Paasonen, P., Asmi, A., Petäjä, T., Kajos, M. K., Äijälä, M., Junninen, H., Holst, T., Abbatt, J.

P. D., Arneth, A., Birmili, W., van der Gon, H. D., Hamed, A., Hoffer, A., Laakso, L., Laaksonen,

A., Leaitch, W. R., Plass-Dülmer, C., Pryor, S. C., Räisänen, P., Swietlicki, E., Wiedensohler, A.,

Worsnop, D. R., Kerminen, V.-M., and Kulmala, M.: Warming-induced increase in aerosol number

concentration likely to moderate climate change, Nature Geoscience, 6, 438–442, 2013.

Paasonen, P., Nieminen, T., Asmi, E., Manninen, H. E., Petäjä, T., Plass-Dülmer, C., Flentje, H.,

Birmili, W., Wiedensohler, A., Hõrrak, U., Metzger, A., Hamed, A., Laaksonen, A., Facchini,

M. C., Kerminen, V.-M., and Kulmala, M.: On the roles of sulphuric acid and low-volatility

organic vapours in the initial steps of atmospheric new particle formation, Atmos. Chem. Phys.,

10, 11223–11242, 2010.

Partanen, A.-I., Kokkola, H., Romakkaniemi, S., Kerminen, V.-M., Lehtinen, K. E. J., Bergman, T.,

Arola, A., and Korhonen, H.: Direct and indirect effects of sea spray geoengineering and the role

of injected particle size, J. Geophys. Res., 117, 2012.

Penner, J. E., Dong, X., and Chen, Y.: Observational evidence of a change in radiative forcing due

to the indirect aerosol effect, Nature, 427, 231–234, 2004.

Petäjä, T., Mauldin, R., Kosciuch, E., McGrath, J., Nieminen, T., Paasonen, P., Boy, M., Adamov,

A., Kotiaho, T., and Kulmala, M.: Sulfuric acid and OH concentrations in a boreal forest site,

Atmos. Chem. Phys., 9, 7435–7448, 2009.


Petäjä, T., Sipilä, M., Paasonen, P., Nieminen, T., Kurtén, T., Ortega, I. K., Stratmann, F.,

Vehkamäki, H., Berndt, T., andKulmala, M.: ExperimentalObservationofStronglyBoundDimers

of Sulfuric Acid: Application to Nucleation in the Atmosphere, Phys. Rev. Lett., 106, 228302, 2011.

Peters, A. and Pope, C. A. I.: Cardiopulmonary mortality and air pollution, Lancet, 360, 1184–1185,


Petters, M. D. and Kreidenweis, S. M.: A single parameter representation of hygroscopic growth and

cloud condensation nucleus activity, Atmos. Chem. Phys., 7, 1961–1971, 2007.

Petters, M. D., Wex, H., Carrico, C. M., Hallbauer, E., Massling, A., McMeeking, G. R., Poulain,

L., Wu, Z., Kreidenweis, S. M., and Stratmann, F.: Towards closing the gap between hygroscopic

growth and activation for secondary organic aerosol – Part 2: Theoretical approaches, Atmos.

Chem. Phys., 9, 3999–4009, 2009.

Pope, C., Burnett, R., Thun, M., Calle, E., Krewski, D., Ito, K., and Thurston, G.: Lung cancer,

cardiopulmonary mortality, and long-term exposure to fine particulate air pollution, JAMA, 287,

1132–1141, 2002.

Pringle, K. J., Tost, H., Pozzer, A., Pöschl, U., and Lelieveld, J.: Global distribution of the effective

aerosol hygroscopicity parameter for CCN activation, Atmos. Chem. Phys., 10, 5241–5255, 2010.

Quinn, P. K., Bates, T. S., Coffman, D. J., and Covert, D. S.: Influence of particle size and chemistry

on the cloud nucleating properties of aerosols, Atmos. Chem. Phys., 8, 1029–1042, 2008.

Ramanathan, V., Ramana, M. V., Roberts, G., Kim, D., Corrigan, C., Chung, C., and Winker, D.:

Warming trends in Asia amplified by brown cloud solar absorption, Nature, 448, 575–U5, 2007.

Rannik, Ü., Aalto, P., Keronen, P., Vesala, T., and Kulmala, M.: Interpretation of aerosol particle

fluxes over a pine forest: Dry deposition and random errors, J. Geophys. Res., 108, 2003.

Reid, J., Koppmann, R., Eck, T., and Eleuterio, D.: A review of biomass burning emissions part II:

intensive physical properties of biomass burning particles, Atmos. Chem. Phys., 5, 799–825, 2005.

Reiner, T. and Arnold, F.: Laboratory investigations of gaseous sulfuric-acid formation via

SO 3 +H 2 O+M→H 2 SO 4 +M – measurement of the rate-constant and product identification, J.

Chem. Phys., 101, 7399–7407, 1994.

Riipinen, I., Pierce, J. R., Yli-Juuti, T., Nieminen, T., Häkkinen, S., Ehn, M., Junninen, H., Lehtipalo,

K., Petäjä, T., Slowik, J., Chang, R., Shantz, N. C., Abbatt, J., Leaitch, W. R., Kerminen, V. M.,

Worsnop, D. R., Pandis, S. N., Donahue, N. M., and Kulmala, M.: Organic condensation: a

vital link connecting aerosol formation to cloud condensation nuclei (CCN) concentrations, Atmos.

Chem. Phys., 11, 3865–3878, 2011.

Riipinen, I., Yli-Juuti, T., Pierce, J. R., Petäjä, T., Worsnop, D. R., Kulmala, M., and Donahue,

N. M.: The contribution of organics to atmospheric nanoparticle growth, Nature Geoscience, 5,

453–458, 2012.

Rohrer, F. and Berresheim, H.: Strong correlation between levels of tropospheric hydroxyl radicals

and solar ultraviolet radiation, Nature, 442, 184–187, 2006.

Romakkaniemi, S., Kokkola, H., Lehtinen, K. E. J., and Laaksonen, A.: The influence of nitric acid

on the cloud processing of aerosol particles, Atmos. Chem. Phys., 6, 1627–1634, 2006.

Salonen, M., Kurtén, T., Vehkamäki, H., Berndt, T., and Kulmala, M.: Computational investigation

ofthe possibleroleofsomeintermediateproductsofSO 2 oxidationin sulfuricacid-waternucleation,

Atmos. Res., 91, 47–52, 2009.


Seinfeld, J. H. and Pandis, S. N.: Atmospheric Chemistry and Physics: From air pollution to climate

change, John Wiley & Sons, Inc., USA, 2006.

Sellegri, K., Hanke, M., Umann, B., Arnold, F., and Kulmala, M.: Measurements of organic gases

during aerosol formation events in the boreal forest atmosphere during QUEST, Atmos. Chem.

Phys., 5, 373–384, 2005.

Sipilä, M., Lehtipalo, K., Attoui, M., Neitola, K., Petäjä, T., Aalto, P. P., O’Dowd, C. D., and Kulmala,

M.: Laboratory verification of PH-CPC’s ability to monitor atmospheric sub-3 nm clusters,

Aerosol Sci. Tech., 43, 126–135, 2009.

Sipilä, M., Berndt, T., Petäjä, T., Brus, D., Vanhanen, J., Stratmann, F., Patokoski, J., Mauldin,

R., Hyvärinen, A., Lihavainen, H., and Kulmala, M.: The role of sulfuric acid in atmospheric

nucleation, Science, 327, 1243–1246, 2010.

Spracklen, D., Carslaw, K., Kulmala, M., Kerminen, V.-M., Sihto, S.-L., Riipinen, I., Merikanto,

J., Mann, G., Chipperfield, M., Wiedensohler, A., Birmili, W., and Lihavainen, H.: Contribution

of particle formation to global cloud condensation nuclei concentrations, Geophys. Res. Lett., 35,

L06808, 2008.

Spracklen, D. V., Pringle, K. J., Carslaw, K. S., Chipperfield, M. P., and Mann, G. W.: A global

off-line model of size-resolved aerosol microphysics: I. Model development and prediction of aerosol

properties, Atmos. Chem. Phys., 5, 2227–2252, 2005.

Spracklen, D. V., Carslaw, K. S., Kulmala, M., Kerminen, V.-M., Mann, G. W., and Sihto, S.-L.: The

contribution of boundary layer nucleation events to total particle concentrations on regional and

global scales, Atmos. Chem. Phys., 6, 5631–5648, 2006.

Stanier, C., Khlystov, A., and Pandis, S.: Nucleation events during the Pittsburgh air quality study:

Description and relation to key meteorological, gas phase, and aerosol parameters, Aerosol Sci.

Tech., 38, 253–264, 2004.

Torpo, L., Kurtén, T., Vehkamäki, H., Laasonen, K., Sundberg, M., and Kulmala, M.: Significance of

ammonia in growth of atmospheric nanoclusters, J. Phys. Chem. A, 111, 10671–10674, 2007.

Tunved, P., Hansson, H., Kerminen, V.-M., Strom, J., Dal Maso, M., Lihavainen, H., Viisanen, Y.,

Aalto, P. P., Komppula, M., and Kulmala, M.: High natural aerosol loading over boreal forests,

Science, 312, 261–263, 2006a.

Tunved, P., Korhonen, H., Strom, J., Hansson, H., Lehtinen, K. E. J., and Kulmala, M.: Is nucleation

capable of explaining observed aerosol integral number increase during southerly transport over

Scandinavia?, Tellus B, 58, 129–140, 2006b.

Tunved, P., Ström, J., Kulmala, M., Kerminen, V.-M., Dal Maso, M., Svenningson, B., Lunder, C.,

and Hansson, H.: The natural aerosol over Northern Europe and its relation to anthropogenic

emissions - implications of important climate feedbacks, Tellus B, 60, 473–484, 2008.

Twomey, S.: Aerosols, clouds and radiation, Atmos. Environ. A, 25, 2435–2442, 1991.

Vakkari, V., Laakso, H., Kulmala, M., Laaksonen, A., Mabaso, D., Molefe, M., Kgabi, N., and Laakso,

L.: New particle formation events in semi-clean South African savannah, Atmos. Chem. Phys., 11,

3333–3346, 2011.

Vanhanen, J., Mikkilä, J., Lehtipalo, K., Sipilä, M., Manninen, H. E., Siivola, E., Petäjä, T., and

Kulmala, M.: Particle Size Magnifier for nano-CN detection, Aerosol Sci. Tech., 45, 533–542, 2011.

Vehkamäki, H.: Classical Nucleation Theory in Multicomponent Systems, Springer, Springer-Verlag,

Berlin Heidelberg, Germany, 2006.


Vehkamäki, H., Kulmala, M., Napari, I., Lehtinen, K. E. J., Timmreck, C., Noppel, M., and Laaksonen,

A.: An improved parameterization for sulfuric acid-water nucleation rates for tropospheric

and stratospheric conditions, J. Geophys. Res., 107, 4622, 2002.

Vehkamäki, H., McGrath, M. J., Kurtén, T., Julin, J., Lehtinen, K. E. J., and Kulmala, M.: Rethinking

the application of the first nucleation theorem to particle formation, J. Chem. Phys., 136,


Viana, M., Kuhlbusch, T., Querol, X., Alastuey, A., Harrison, R., Hopke, P., Winiwarter, W., Vallius,

M., Szidat, S., Prévôt, A., Hueglin, C., Bloemen, H., Wåhlin, P., Vecchi, R., Miranda, A., Kasper-

Giebl, A., Maenhaut, W., and Hitzenberger, R.: Source apportionment of particulate matter in

Europe: A review of methods and results, J. Aerosol Sci., 39, 827–849, 2008.

Viisanen, Y., Kulmala, M., and Laaksonen, A.: Experiments on gas-liquid nucleation of sulfuric acid

and water, J. Chem. Phys., 107, 920–926, 1997.

Virtanen, A., Joutsensaari, J., Koop, T., Kannosto, J., Yli-Pirilä, P., Leskinen, J., Mäkelä, J.,

Holopainen, J., Pöschl, U., Kulmala, M., et al.: An amorphous solid state of biogenic secondary

organic aerosol particles, Nature, 467, 824–827, 2010.

Vuollekoski, H., Kerminen, V.-M., Anttila, T., Sihto, S.-L., Vana, M., Ehn, M., Korhonen, H., McFiggans,

G., O’Dowd, C. D., and Kulmala, M.: Iodine dioxide nucleation simulations in coastal and

remote marine environments, J. Geophys. Res., 114, 2009.

W. Birmili and A. Wiedensohler: New particle formation in the continental boundary layer: Meteorological

and gas phase parameter influence, Geophys. Res. Lett., 27, 3325–3328, 2000.

Wang, L., Qi, J. H., Shi, J. H., Chen, X. J., and Gao, H. W.: Source apportionment of particulate

pollutants in the atmosphere over the Northern Yellow Sea, Atmos. Environ., 70, 425–434, 2013.

Wang, T., Nie, W., Gao, J., Xue, L. K., Gao, X. M., Wang, X. F., Qiu, J., Poon, C. N., Meinardi,

S., Blake, D., Wang, S. L., Ding, A. J., Chai, F. H., Zhang, Q. Z., and Wang, W. X.: Air quality

during the 2008 Beijing Olympics: secondary pollutants and regional impact, Atmos. Chem. Phys.,

10, 7603–7615, 2010.

Weber, R., Marti, J., McMurry, P., Eisele, F., Tanner, D., and Jefferson, A.: Measured atmospheric

new particle formation rates: Implications for nucleation mechanisms, Chem. Eng. Comm., 151,

53–64, 1996.

Weber, R., Marti, J., McMurry, P., Eisele, F., Tanner, D., and Jefferson, A.: Measurements of new

particle formation and ultrafine particle growth rates at a clean continental site, J. Geophys. Res.,

102, 4375–4385, 1997.

Weber, R., McMurry, P., Eisele, F., and Tanner, D.: Measurement of expected nucleation precursor

species and 3-500-nm diameter particles at Mauna-Loa-Observatory, Hawaii, J. Atmos. Sci., 52,

2242–2257, 1995.

Welz, O., Savee, J. D., Osborn, D. L., Vasu, S. S., Percival, C. J., Shallcross, D. E., and Taatjes, C. A.:

Direct Kinetic Measurements of Criegee Intermediate (CH 2 OO) Formed by Reaction of CH 2 I with

O 2 , Science, 335, 204–207, 2012.

Wiedensohler, A., Cheng, Y. F., Nowak, A., Wehner, B., Achtert, P., Berghof, M., Birmili, W., Wu,

Z. J., Hu, M., Zhu, T., Takegawa, N., Kita, K., Kondo, Y., Lou, S. R., Hofzumahaus, A., Holland,

F., Wahner, A., Gunthe, S. S., Rose, D., Su, H., and Poeschl, U.: Rapid aerosolparticle growth and

increase of cloud condensation nucleus activity by secondary aerosol formation and condensation:

A case study for regional air pollution in northeastern China, J. Geophys. Res., 114, 2009.


Williams, J., de Reus, M., Krejci, R., Fischer, H., and Ström, J.: Application of the variability-size

relationship to atmospheric aerosol studies: estimating aerosol lifetimes and ages, Atmos. Chem.

Phys., 2, 133–145, 2002.

Wu, Z., Hu, M., Lin, P., Liu, S., Wehner, B., and Wiedensohler, A.: Particle number size distribution

in the urban atmosphere of Beijing, China, Atmospheric Environment, 42, 7967–7980, 2008.

Yli-Juuti, T., Nieminen, T., Hirsikko, A., Aalto, P. P., Asmi, E., Hõrrak, U., Manninen, H. E.,

Patokoski, J., Dal Maso, M., Petäjä, T., Rinne, J., Kulmala, M., and Riipinen, I.: Growth rates of

nucleation mode particles in Hyytiala during 2003-2009: variation with particle size, season, data

analysis method and ambient conditions, Atmos. Chem. Phys., 11, 12865–12886, 2011.

Young, L., Benson, D., Kameel, F., Pierce, J., Junninen, H., Kulmala, M., and Lee, S.: Laboratory

studies of H 2 SO 4 /H 2 O binary homogeneous nucleation from the SO 2 +OH reaction: evaluation of

the experimental setup and preliminary results, Atmos. Chem. Phys., 8, 4997–5016, 2008.

Yu, F. and Turco, R.: Ultrafine aerosol formation via ion-mediated nucleation, Geophys. Res. Lett.,

27, 883–886, 2000.

Yu, F., Wang, Z., Luo, G., and Turco, R.: Ion-mediated nucleation as an important global source of

tropospheric aerosols, Atmos. Chem. Phys., 8, 2537–2554, 2008.

Yu, F., Luo, G., Bates, T. S., Anderson, B., Clarke, A., Kapustin, V., Yantosca, R. M., Wang,

Y., and Wu, S.: Spatial distributions of particle number concentrations in the global troposphere:

Simulations, observations, and implicationsfor nucleationmechanisms, J. Geophys. Res., 115, 2010.

Yu, H., Kaufman, Y., Chin, M., Feingold, G., Remer, L., Anderson, T., Balkanski, Y., Bellouin, N.,

Boucher, O., Christopher, S., DeCola, P., Kahn, R., Koch, D., Loeb, N., Reddy, M., Schulz, M.,

Takemura, T., and Zhou, M.: A review of measurement-based assessments of the aerosol direct

radiative effect and forcing, Atmos. Chem. Phys., 6, 613–666, 2006.

Yue, D. L., Hu, M., Zhang, R. Y., Wang, Z. B., Zheng, J., Wu, Z. J., Wiedensohler, A., He, L. Y.,

Huang, X. F., and Zhu, T.: The roles of sulfuric acid in new particle formation and growth in the

mega-city of Beijing, Atmos. Chem. Phys., 10, 4953–4960, 2010.

Zhang, Q., Jimenez, J. L., Canagaratna, M. R., Ulbrich, I. M., Ng, N. L., Worsnop, D. R., and Sun,

Y.: Understanding atmospheric organic aerosols via factor analysis of aerosol mass spectrometry:

A review, Anal. Bioanal. Chem., 401, 3045–3067, 2011.

Zhang, R., Suh, I., Zhao, J., Zhang, D., Fortner, E. C., Tie, X., Molina, L. T., and Molina, M. J.:

Atmospheric new particle formation enhanced by organic acids, Science, 304, 1487–1490, 2004.

Zhao, J., Smith, J. N., Eisele, F. L., Chen, M., Kuang, C., and McMurry, P. H.: Observation of

neutral sulfuric acid-amine containing clusters in laboratory and ambient measurements, Atmos.

Chem. Phys., 11, 10823–10836, 2011.

Zhao, J., Eisele, F. L., Titcombe, M., Kuang, C., and McMurry, P. H.: Chemical ionization mass

spectrometric measurements of atmospheric neutral clusters using the cluster-CIMS, J. Geophys.

Res., 115, 2010.


More magazines by this user
Similar magazines